Autonomous coverage robot navigation system

This application is the name "ROBOT NETWORKING, THEMING AND COMMUNICATION SYSTEM" filed on December 2, 2005, and is assigned to the US provisional patent application No. 60 / 741,442, which is subject to 35 USC 119. Claim priority under Article (e). The entire contents of the above application are incorporated herein by reference.

The present invention relates to robots, and more specifically to autonomous coverage robots and related navigation systems.

An autonomous robot is a robot that allows a human to perform a desired task in an unstructured environment without continuous guidance. Many types of robots are to some extent autonomous. Other robots can be made autonomous in different ways. Autonomous coverage robots traverse work surfaces without humans continuously guiding them to perform one or more tasks. In the field of home, office, and / or consumer-oriented robotics, mobile robots that perform domestic roles such as vacuum cleaner cleaning, floor cleaning, patrols, lawnmowers, and other such tasks are widespread. It is used.

<p> On one side, the autonomous coverage robot is located in the front part of the pedestal, with a pedestal that determines the forward drive direction, a controller supported by the pedestal, an omnidirectional receiver supported by the pedestal, and a pedestal from the drive direction. Includes a directional receiver that corresponds to the incidence of emissions on the anterior portion of the. The directional receiver is provided to the mounting receptacle that defines the first and second openings that are approximately aligned with the drive direction, and to the emissions that are housed in the receptacle and received through the first and second openings, respectively. Includes first and second component receivers, respectively arranged correspondingly. Each component receiver is configured to generate its own signal in response to emissions. The controller is configured to determine the direction of emission with respect to the drive direction based on the signal generated by the component receiver.</p><p> The directional receiver is built onto the front part of the pedestal and housed within the mounting receptacle. The mounting structure secures and supports directional receivers used for docking and navigation throughout the room. A directional receiver includes two detectors that are highly directional in the sense that they can detect overlapping signal peaks from a distance, and a servo that involves the source of the beam using those peaks. In one embodiment, the first and second component receivers of the directional receiver are arranged so as to form an angle of about 1-15 ° between them. In addition, the first and second component receivers are collimated so that they are in focus approximately 3-5 meters from the robot, or at any position within their respective bounded areas.</p><p> The controller moves the robot to detect emissions by an omnidirectional receiver, and in response to the detection, orients the pedestal based on the signal generated by the component receiver of the directional receiver. Align the drive direction with the determined emission direction. The controller orients the pedestal based on the signal generated by the component receiver of the directional receiver in response to the detection of the emission by the directional receiver, and the direction of the emission determined to be the robot drive direction. And align. The receiver can be configured to receive the transmission of infrared light.</p><p> In one embodiment, the omnidirectional receiver has an upper portion and includes a housing defining an internal cavity, a cone, and an emission receiver. This upper portion allows emissions to be transmitted to the internal cavity. The conical reflector is placed on top of the cavity to reflect emissions to the upper portion of the housing into the cavity. The emission receiver is located in the cavity below the conical reflector.</p><p> In one embodiment, the robot includes a floor cleaning assembly supported by a pedestal, a cleaning bin supported by the pedestal and arranged to collect debris removed from the floor by the cleaning assembly, and a bin sensor. The bin sensor is configured to indicate that the cleaning bin has reached the filling threshold and to generate a signal to the controller to initiate a docking sequence.</p><p> In another aspect, autonomous mobile robot systems for bounded domains include navigation beas and autonomous coverage robots. The navigation beacon has a gateway beacon emitter arranged to transmit gateway marking emissions while being arranged within the gateway between the first bounded area and the adjacent second bounded area. The autonomous coverage robot is a drive system configured to move the robot around the first bounded area in a cleaning mode in which the robot changes direction in response to the detection of gateway marking emissions and a beacon emission sensor that supports beacon emissions. And include. The drive system is also configured to move the robot through a gateway into a second bounded area in migration mode.</p><p> The beacon can be configured to emit an infrared signal for gateway marking emissions, or to emit other types of signals, so that this signal does not "flow" into the adjacent room. It is also easily blocked by a wall to uniquely identify where the current room or robot is located.</p><p> In one embodiment, the robot remains in the first bounded area when the robot changes direction in response to detection of gateway marking emissions in cleaning mode. The gateway marking emissions can be infrared light and the beacon emission sensor is configured to detect the transmission of infrared light.</p><p> In another embodiment, the drive system is configured to move the robot around the first bounded area in cleaning mode for a predetermined time interval and automatically start migration mode at the end of the time interval. .. The drive system is configured to move the robot in crossing the gateway emissions in migration mode.</p><p> In yet another embodiment, the drive system causes the robot to move around the first bounded area in cleaning mode until a predetermined number of gateway marking emission contacts have been detected, and then automatically enter migration mode. It is composed of. The dwell time of the coverage robot in one room or selective room is the timer or schedule, the number of collisions or mutual interferences with the proximity beam of a particular beacon, the features or number of dust or debris, the battery level, And can also be set by maintenance or remote control override.</p><p> In one embodiment, the robot comprises a floor cleaning system that removes debris from the floor when the robot is moved.</p><p> In some embodiments, the robot comprises a transmitter that signals a beacon to interrupt the transmission of gateway emissions in migration mode.</p><p> In another embodiment, the navigation beacon comprises a vectoring beacon emitter, which, while located within the gateway, is arranged to transmit directed vectoring emissions to a first bounded region. The robot drive system is configured to guide the robot towards the beacon when it encounters vectoring emissions in migration mode. The navigation beacon transmits directed vectoring emissions (which can be infrared light) at an angle adjacent to the cross-gateway direction of approximately 45-90 °. During migration mode, the robot identifies the position of the navigation beacon in response to the detection of directed vectoring emissions, and the drive defined by the robot with respect to the emission path defined by the directed vectoring emissions. Align the directions. The robot advances along the emission path to reach and cross the gateway. The navigation beacon can also transmit proximity emissions laterally around the beacon, allowing the robot to avoid cleaning and migration within the proximity emissions.</p><p> In some embodiments, the robot wirelessly communicates with the navigation beacon to stop the directed vectoring emissions during cleaning mode and activate the directed vectoring emissions during migration. In another embodiment, when the robot enters migration mode, it remotely activates the directed vectoring emissions of the navigation beacon to stop the gateway emissions. In addition, when the robot exits the migration mode, it remotely stops the directed vectoring emissions of the navigation beacon and activates the gateway emissions. The robot activates and stops the emission of the navigation beacon by radio frequency communication. The navigation beacon can also be configured to communicate schedule information to the robot.</p><p> The system can also include a base station located in one of the bounded areas. The base station includes a base that determines the docking direction in which the robot can be properly docked, and a charger for the robot housed within the base. The omnidirectional beam emitter can be mounted on a base and configured to project a proximity beam laterally around the docking station. Two navigation field emitters are housed in the base, each configured to radiate field emissions of overlapping signal beams laterally bounded. One of the radial fields is aligned with the docking direction and defines a lateral field edge that overlaps the other fields. The two navigation field emitters of the base station are arranged so as to form an angle of about 45-90 ° between them. The field of emission of the two navigation field emitters of the base station can be infrared light.</p><p> In another aspect, a method of navigating an autonomous coverage robot between bounded domains involves placing a navigation beacon within a gateway between adjacent first and second bounded domains. The beacon is configured to cross the gateway and carry gateway marking emissions. In some embodiments, the navigation beacon can also transmit proximity emissions laterally around the beacon, allowing the robot to avoid cleaning and migration within the proximity emissions. The method also includes the step of placing the coverage robot within the first bounded area. The robot autonomously traverses the first bounded area in cleaning mode, and when it encounters a gateway marking emission within the gateway, it remains in the first bounded area, thereby moving the robot to the second area. Avoid migration. When the cleaning mode ends in the first area, the robot autonomously initiates the migration mode, moves through the gateway, passes through the beacon, and enters the second bounded area.</p><p> In some embodiments, the robot remotely stops the gateway marking emissions of the navigation bea when it enters the migration mode and activates the gateway marking emissions when it exits the migration mode. The gateway marking emission can be infrared light. In another embodiment, the robot does not respond to gateway emissions during migration mode.</p><p> In one embodiment, when the migration mode ends in the second area, the robot autonomously starts the cleaning mode in the second bounded area.</p><p> In one embodiment, the navigation beacon is configured to transmit directed vectoring emissions to a first bounded region by a beacon located within the gateway. The robot is driven towards the beacon when it detects vectoring emissions in migration mode. The step of detecting the vectoring emission directed in the migration mode is the step of detecting the vectoring emission directed by the robot by the directional receiver on the robot aligned with the driving direction of the robot, and the robot. Includes a step of aligning the driving direction of the robot with the path defined by the directed vectoring emissions. The robot can also detect vectoring emissions directed by an omnidirectional receiver on the robot before being moved to detect vectoring emissions directed by the directional receiver on the robot. Is. The robot travels along the emission path towards the beacon, detects the beacon perimeter emitted by the beacon, and travels along the beam perimeter into a second region through the gateway. By doing so, it passes through the beacon.</p><p> In another aspect, a method of navigating an autonomous coverage robot between bounded domains involves placing a navigation beacon within a gateway between adjacent first and second bounded domains. The beacon is configured to transmit directed emissions to the first bounded region and gateway emissions in the cross-gateway direction. The method also includes the step of placing the coverage robot within the first bounded area. When the robot autonomously traverses the first bounded area in cleaning mode and encounters gateway emissions within the gateway, the robot remains in the first bounded area, thereby moving to the second area. Avoid robot migration. When the cleaning mode ends in the first area, the robot autonomously enters the migration mode and moves to the second bounded area through the gateway by detecting the directed emissions. Then, in response to the directed emissions, it passes through the beacon through the gateway and enters the second area.</p><p> The navigation beacon contains a base that can be placed within the gateway that directs the cross-gateway. The gateway beam emitter is housed in the base and is arranged to radiate the beam in the direction of the cross gateway. The first and second directional beam emitters are housed in the base and are arranged to radiate their respective focal beams when the base is placed in the gateway. The omnidirectional beam emitter is located within the base and is configured to project a proximity beam laterally around the beacon. The first and second directional focal beams form an angle of approximately 45-90 ° in the direction of adjacent cross-gateways.</p><p> Another embodiment of the navigation beacon emits a beam along the line of the "virtual wall" and a circular proximity beam (and / or RF zone) to prevent the robot from colliding when detected by the robot. Includes beam emitter. The first beam emitter can be placed within the gateway so that it intersects the gateway and emits a beam. The second beam emitter can be placed within the gateway to radiate an angled beam to the first room and modulated to indicate the first room. A third beam emitter can be placed within the gateway to radiate an angled beam to a second adjacent room and modulated to indicate a second room.</p><p> The navigation beacon can also act as a virtual, temporary confinement wall or gateway, where the robot stays within the virtual gateway, for example for a given number of virtual gateways interfering or time spans, and then intersects the virtual gateways. It then resets its time or incident and pauses within the next "room" that can be partially or completely defined by the virtual gateway.</p><p> A plurality of beacons can be used to demarcate the boundaries of each bounded region. Directed vectoring and gateway emissions can be infrared light. Each transmission emission is distinguished by a time division multiplexing of random offsets over a period of time. Each beacon is individually encoded (eg, left and right beams) and can be configured by a DIP switch or other device of interest located on the beacon. Beacon coding can be set or changed by the robot (eg, by RF commands issued by the robot) when the robot encounters a contiguous beacon or two similarly encoded beacons. .. Beacons can be set on virtual walls and / or threshold markers. When the robot encounters each beacon, the robot counts the beas and identifies them by beacon modulation. The first one encountered is number one and is considered to be the closest to the base. The second one is the second one, which is supposed to be farther away. The third is number 3, and is supposed to be farther away. The left and right side beams or oriented beams of the beacon are individually encoded. As a result, the robot can record which side is considered closer to the base. When the robot is in a state of returning to the base (eg, timeout, battery drain, bin filling, scheduled end, maintenance required, general event occurrence, or other circumstances), the robot is on the remote side. The beam of the robot searches for the lowest numbered beacon and heads in that direction. The direction of the beam also allows the robot to move properly past the beacon. For example, when leaving the base station, the robot sequentially passes through the directed beam A and the directed beam B of the beacon on either or both of the robot's left sides. Robot to return to base station Waits / searches / cleans until it detects the directed beam B of the beacon. The robot then approaches the beacon by following beam B until it detects the beacon's proximity beam. The robot turns left and follows a curved path along the edge of the proximity beam, searching / cleaning until the robot detects beam A. The robot follows beam A away from the beacon and then searches / cleans the base station. This is just one example of a navigation method that uses the orientation of the left and right beacons to identify the room.</p><p> In some embodiments, when the migration mode ends in the second region, the robot autonomously enters the docking mode and is moved towards the station in the second region. The base station has a base that determines the docking direction in which the robot can be docked properly, a charger for the robot housed in the base, and a proximity beam placed on the base laterally around the base station. An omnidirectional beam emitter configured to project to, and two navigation units housed in the base, each radiating field emissions of laterally bounded and overlapping signal beams. Includes field emitters and. One of the radial fields is aligned with the docking direction and defines a lateral field edge that overlaps the other fields. The robot is moved towards the base station by detecting and advancing along the lateral field edge of the overlapping field aligned with the docking direction until docked at the station. Two navigation field emitters (which can be infrared emitters) are placed at an angle of about 45-90 ° between them. The robot is driven by an omnidirectional receiver on the robot to detect base station emissions and detect the outer lateral field edge of at least one field emission. The robot advances along the outer lateral field edge of one field emission to the aligned lateral field edge of the overlapping fields. Upon detecting an aligned lateral field edge, the robot advances along the aligned lateral field edge until it docks with the station.</p>

<figref num="1A">FIG. 1A is a perspective view showing an embodiment of an autonomous coverage robot.</figref><figref num="1B">FIG. 1B is an exploded view of the autonomous coverage robot.</figref><figref num="1C">Figure 1C is an exploded view of the bumper of the autonomous coverage robot.</figref><figref num="2">FIG. 2 is a diagram showing the positions of the omnidirectional receiver and the directional receiver on the bumper of the autonomous coverage robot.</figref><figref num="3">FIG. 3 is a perspective view of the directional receiver.</figref><figref num="4">FIG. 4 is a front view of the directional receiver.</figref><figref num="5">FIG. 5 is an exploded view of the directional receiver.</figref><figref num="6">FIG. 6 is a cut-out view of the omnidirectional receiver.</figref><figref num="7">FIG. 7 is a side view of the omnidirectional receiver.</figref><figref num="8">FIG. 8 is a perspective view of the omnidirectional receiver.</figref><figref num="9">FIG. 9 is a perspective view of the lower part of the autonomous coverage robot.</figref><figref num="10">FIG. 10 is an exploded view of an exemplary navigation beacon.</figref><figref num="11">FIG. 11 is a perspective view and a front view of an exemplary beacon emitter assembly.</figref><figref num="12">FIG. 12 is a perspective view and a front view of an exemplary beacon emitter assembly.</figref><figref num="13">FIG. 13 is a diagram illustrating an embodiment of a simplified navigation beacon that can be used with an autonomous mobile robot navigation system.</figref><figref num="14">FIG. 14 is a variety of exemplary base stations that can be included within an autonomous mobile robot system.</figref><figref num="15">FIG. 15 is a variety of exemplary base stations that can be included within an autonomous mobile robot system.</figref><figref num="16">FIG. 16 is a variety of exemplary base stations that can be included within an autonomous mobile robot system.</figref><figref num="17">FIG. 17 is a variety of exemplary base stations that can be included within an autonomous mobile robot system.</figref><figref num="18">FIG. 18 is a perspective view of an exemplary emitter assembly used within a base station.</figref><figref num="19">FIG. 19 is a top view of an exemplary emitter assembly used within a base station.</figref><figref num="20">FIG. 20 is a rear view of an exemplary emitter assembly used within a base station.</figref><figref num="21">FIG. 21 is a side view of an exemplary emitter assembly used within a base station.</figref><figref num="22">FIG. 22 is a block diagram of a typical electronic component structure of an autonomous coverage robot.</figref><figref num="23A">FIG. 23A is a diagram showing a first embodiment in which a method of navigating an autonomous coverage robot between bounded regions is applied by adding a base station and using a navigation beacon.</figref><figref num="23B">FIG. 23B is a diagram showing a second embodiment in which two or more beams are encountered at the same time.</figref><figref num="24A">FIG. 24A is a block diagram showing the software architecture for navigating an autonomous coverage robot.</figref><figref num="24B">FIG. 24B is a block diagram showing the software architecture for navigating an autonomous coverage robot.</figref><figref num="25A">FIG. 25A is a schematic diagram of a robot moving from one area to another using a navigation beacon.</figref><figref num="25B">FIG. 25B is a schematic diagram of a robot moving from one area to another using a navigation beacon.</figref><figref num="25C">FIG. 25C is a schematic diagram of a robot moving from one area to another using a navigation beacon.</figref><figref num="26">FIG. 26 is a schematic diagram of the beam alignment behavior of the robot.</figref><figref num="27">FIG. 27 is a schematic diagram of the robot's oriented beam homing behavior.</figref><figref num="28">FIG. 28 is a schematic diagram of the field alignment behavior of the robot.</figref><figref num="29">FIG. 29 is a schematic diagram of the field-following behavior of the robot.</figref><figref num="30">FIG. 30 is a schematic diagram of the beacon withdrawal behavior of the robot.</figref><figref num="31A">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="31B">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="31C">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="31D">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="31E">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="31F">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="31G">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="31H">31A-31H are schematic views of a robot moving from one area to another using a navigation beacon.</figref><figref num="32A">32A to 32E are schematic views of the robot docking with the base station.</figref><figref num="32B">32A to 32E are schematic views of the robot docking with the base station.</figref><figref num="32C">32A to 32E are schematic views of the robot docking with the base station.</figref><figref num="32D">32A to 32E are schematic views of the robot docking with the base station.</figref><figref num="32E">32A to 32E are schematic views of the robot docking with the base station.</figref>

The same reference numerals in the various drawings indicate the same elements.

1A and 1B are an upper partial perspective view and an exploded view of the autonomous coverage robot 100. The robot 100 has a pedestal 102, a controller 103, an omnidirectional receiver 104, and a directional receiver 106. The pedestal 102 has a forward drive direction and carries the controller 103 and the receivers 104 and 106 on the bumper 107. Receivers 104 and 106 provide navigation information to controller 103. Using the inputs from receivers 104 and 106, controller 103 raises a command to be executed by robot 100. As a result, the robot 100 can perform floor cleaning in an autonomous manner.

FIG. 2 is a diagram showing the positions of the omnidirectional receiver 104 and the directional receiver 106 on the bumper 107 of the robot 100. The bumper 107 can also have other sensors to assist the robot 100 as it navigates around it, as shown in FIG. 1C. The proximity sensor 1072 can be used to determine if an obstacle is near the robot 100. The contact sensor 1074 can be used to determine that the robot 100 has physically encountered an object. The cliff sensor 1076 can be used to detect when the robot 100 encounters an edge of the floor, such as when it encounters a series of stairs.

The omnidirectional sensor 104 can be used to detect that the robot 100 is in the vicinity of the navigation beacon. For example, the omnidirectional sensor 104 can relay a signal indicating the intensity of emission to the control system, and the stronger the signal, the closer it is to the navigation beacon.

The proximity sensor 1072 can be used to detect nearby obstacles. Proximity sensor 1072 can be, for example, an infrared or ultrasonic sensor that provides a signal when an object is within a given range of robot 100.

A contactor or bump sensor 1074 can be used to detect whether the robot 100 has physically encountered an obstacle. Such a sensor 1074 can use physical properties such as capacitance or physical movement within the robot 100 to determine that the robot has encountered an obstacle.

The cliff sensor 1074 can detect that the robot 100 has encountered a floor edge, such as when it encounters a series of stairs. When an edge is detected, the robot 100 can have an action such as changing the moving direction of the robot.

In some embodiments, the cliff sensor 1074 can be incorporated into the mounting device, which secures and protects the sensor so that the sensor points towards a window built into the bottom of the mounting device. Deploy. Along with this sensor, a mounting device and a window constitute a cliff sensor unit. For example, four cliff sensor units can be incorporated in the bumper.

The reliability of the cliff sensor can be increased by reducing the accumulation of dust. In some embodiments, the window can be incorporated into the bottom of the mounting device, which is placed in a slanted molded body made of a material that prevents the buildup of dust, such as antistatic material. Includes shield. The shield components and moldings can be welded together. In order to further reduce the accumulation of dust and dirt, it is possible to place the shield on an inclined surface to make it easier to slide off the dust.

In some embodiments, a secondary cliff sensor can be provided behind the existing cliff sensor to detect the edge of the floor if the wheel drop sensor on the caster wheel fails.

The controller 103 can be configured to propel the robot 100 based on the directional and velocity settings. The signals received from the proximity and contact sensor can be used by the control system to issue commands to deal with obstacles. Signals from proximity and contact sensors allow the control system to change the commanded speed or direction of the robot 100. For example, a signal from a proximity sensor caused by an adjacent wall can cause the control system to issue a deceleration command. In another case, the collision signal from the contact sensor due to contact with the chair can cause the control system to issue a command to change direction. In other cases, the robot's speed setting may not be able to be lowered in response to the contact sensor, and / or the robot's orientation setting may not be able to be changed in response to the proximity sensor, but instead , It is possible to operate the two sensors independently.

The controller 103 can include a first independent action routine configured to adjust the speed setting of the robot and a second independent action routine configured to change the direction setting of the robot. Yes, the first and second independent action routines are configured to run in parallel and independently of each other. The first independent action routine can be configured to poll the proximity sensor and the second independent action routine can be configured to poll the dynamic bump sensor.

3 to 5 are a perspective view, a front view, and an exploded view of the directional receiver 104. The directional receiver 104 is arranged in front of the pedestal 102. Emissions can be received by the directional receiver 104 along the drive direction, which generates the corresponding signal and sends it to the controller 103. The directional receiver 104 includes a mounting receptacle 108 with a first opening 110 and a second opening 112 aligned with the drive direction of the pedestal 102. The openings 110 and 112 include a first component receiver 114 and a second component receiver 116.

The components receivers 114 and 116 are such that emissions that occur along the drive direction and are directed at the directional receiver 104 are directed through the openings 110 and 112 to the component receivers 114 and 116, respectively. Is located relative to the openings 110 and 112. Emissions directed at the component receivers 114 and 116 generate a corresponding signal, which is used by controller 103 to determine the direction of the emission with respect to the driving direction of the robot. The cover 117 is mounted on the upper portion of the mounting receptacle 108 to prevent emissions from pointing towards the component receivers 114 and 116 along the drive direction. In some embodiments, it is possible to secure the cover 117 to the mounting receptacle 108 using a lock tab 119 or other fastening method.

In some embodiments, the components receivers 114 and 116 of the directional receiver 104 are capable of detecting the homing beam with a substantially 10% to 30% overlap. In one embodiment, the first and second component receivers 114 and 116 can be arranged so that the emissions directed at them make an angle of 1-15 °. In another embodiment, the first and second component receivers 114 and 116 can be aligned so that the emissions aligned with them intersect at 3-5 meters in front of the robot. In one embodiment, the respective component receivers 114 and 116 have a cleaning center of about 14 ° from the normal direction of the directional receiver 104, a spread angle of about 28-33 °, and other component receivers. Emissions are received within a cleaning range that overlaps 114 and 116 approximately 10 °.

In some cases, the controller 103 can move the robot 100 to detect emissions from the omnidirectional receiver 104. The direction of emission can be determined using the component receivers 114 and 116. The controller 103 can orient the pedestal 102 in response to the detection to align the robot drive direction with the determined emission direction.

In other cases, the controller 103 orients the pedestal 102 with the driving direction of the robot based on the signals generated by the component receivers 114 and 116 in response to the emissions detected by the directional receiver 106. It is possible to align with the determined emission direction.

6 to 8 are a perspective view, a side view, and a cut-out view of the omnidirectional receiver 104. The omnidirectional receiver 104 can include a housing 118, a conical reflector 120, and an emission receiver 122. Housing 118 has an upper portion 124 and an internal cavity 126. The upper portion 124 allows emissions to be transmitted to the internal cavity 126. The conical reflector 120 is located on the top surface of the cavity 126 to reflect emissions directed from the upper portion 124 of the housing 118 to the internal cavity 126. The emission receiver 122 is located below the conical reflector 120 in the internal cavity 126.

In some embodiments, receivers 114, 116, and 122 can be configured to receive infrared (IR) transmissions. In such cases, the guide 128 (eg, the light pipe) can guide the emissions reflected by the conical reflector 120 and point them at the emission receiver 122.

FIG. 9 is a perspective view of the lower portion of the autonomous coverage robot 100. The drive system 130 includes a first drive wheel 132 and a second drive wheel 134 that support the pedestal 102. The caster 136 can provide additional support for the pedestal 102. The motor can be mechanically connected to the drive wheels to propel the robot 100, providing forward, reverse, and turning capabilities.

Robot 100 can have a floor cleaning system that removes dust and debris from the floor when moved around it. The floor cleaning system can include a floor cleaning assembly 140, a cleaning bin 142, and a bin sensor (not shown). The floor cleaning assemblies 140 and 146, and the cleaning bin 142 can be supported by the pedestal 102. The cleaning bin 142 can be arranged to collect debris removed from the floor being cleaned by the cleaning assembly 140. The bin sensor can be configured to generate a signal indicating whether the cleaning bin has been filled to a threshold. When the threshold is reached, the controller 103 is capable of initiating a docking sequence (described below) with the base station. The cleaning bin 142 can be manually emptied of its contents, or in some embodiments, the robot 100 can automatically empty the cleaning bin 142 when docked.

In addition to providing the framework for the other elements that make up the robot 100, the pedestal 102 can have a ridge in front of it that is higher than all other positions on the pedestal 102. Such a ridge can stop the robot 100 in the event of encountering an obstacle at a high altitude such as a sofa so that it does not get stuck under it. The controller 103 can typically command the drive system 130 to operate at less than maximum torque (eg, 50% of capacity) so that the robot itself can be released more easily when the robot 100 gets stuck. It is possible. The fit of the robot 100 is detected, for example, by an increase in the current to the drive motor, in which case the controller 103 can increase the torque to release the robot 100.

Another embodiment of the anti-jamming system includes a spring-suspended wheel system with a potentiometer to measure how much the robot 100 is pushed down by a potentially jammed obstacle. Another embodiment includes an infrared sensor located on the underside of the robot 100, which is used to measure the distance the robot 100 is depressed.

FIG. 10 is an exploded view of an exemplary navigation beacon 150. The navigation beacon 150 can be used with existing walls or other obstacles to create bounded areas. The bounded area can be, for example, such that the robot restricts access to the area. The navigation beacon 150, together with the robot 100 described above, forms an embodiment of an autonomous mobile robot system.

The navigation beacon 150 has a beacon emitter assembly 152, which assembly includes a gateway beacon emitter 154 and an omnidirectional emitter 160. The gateway beacon emitter 154 can be arranged to carry gateway marking emissions. For example, the Beacon 150 can be located within a gateway, which separates the first and second adjacent regions and radiates gateway marking emissions to form a boundary. The omnidirectional receiver 104 and the directional receiver 106 on the robot 100 can detect gateway marking emissions, thereby acting as a beacon emission sensor. For example, Beacon 150 and Robot 100 can use infrared (IR) emitters and sensors to create and detect gateway marking emissions.

In one embodiment, the robot 100 controls the state of the navigation beam by means of commands transmitted over a packet wireless network. The address to which the beacon reacts on this network is a combination of a robot address and a node address. After incorporating the battery within the Beacon 150, the Beacon 150 periodically attempts to contact any robot to see if its emitter should be waked up and operated. Robot 100 can react by attracting connections to its network and transmitting radio packets containing temporary addresses to use. While operating by a temporary address, the Beacon 150 forces an infrared code from emitter 154 with a fence beam and from omnidirectional emitter 160 to indicate that it is uncoupled, i.e. its radio address is temporary. It is transmitted by a field beam. When the robot 100 recognizes a beam having a coupling code, it repeatedly transmits a radio packet for each recently assigned temporary address, and transmits a new code with a beam called a wink. When the robot 100 recognizes the wink code, the robot transmits a radio packet containing a new node address to be used thereafter and a time indicating the number of hours when the use of the address is directed. By successfully coupling to Robot 100, Beacon 150 will wake up only when Robot 100 is present. The wireless communication confirmation technology using infrared light signals is designed so that the robot 100, which does not exist on the same floor as the beacon 150, is not permanently controlled.

The drive system 130 can be configured to move the robot 100 around the first region during the cleaning mode. In the cleaning mode, the robot 100 can change direction in response to the detection of gateway marking emissions. In addition, the drive system 130 can also move the robot 100 to a second bounded region via a gateway while in migration mode.

During the cleaning mode, the robot 100 can remain in the first bounded region by changing its drive direction when it detects a gateway marking emission. Therefore, the gateway marking emissions act as a virtual barrier that can prevent the robot 100 from leaving the first bounded area.

In some cases, the drive system 130 can be configured to move the robot 100 around the first bounded area in cleaning mode for a preset time interval. The drive system 130 can be automatically put into migration mode when a preset time interval has elapsed. During migration mode, drive system 130 can be configured to move robot 100 across gateway emissions. In other cases, the drive system can be configured to move the robot around the first bounded area in cleaning mode until it encounters a preset number of gateway marking emissions. It is possible to automatically start the migration mode when the gateway marking emission is encountered a preset number of times.

Robot 100 can include a transmitter for communicating with Beacon 150. A transmitter can be used to signal Beacon 150 to stop or suspend the transmission of gateway emissions in migration mode. By signaling the Beacon 150 to turn on various emitters only when needed, the system can implement power saving features. Such a function can serve to extend the battery life of the Beacon 150.

11 and 12 are perspective views and front views of the Beacon Emitter Assembly 152. Beacon Emitter Assembly 152 includes a first oriented vectoring Beacon Emitter 156 and a second directed Vectoring Beacon Emitter 158. Directed vectoring beam emitters can be used to create emission fields with characteristic edges and diffusion patterns that can be used to define navigation routes. The navigation beacon 150 can be located within a gateway between two bounded regions with a vectoring beacon emitter 156 arranged to transmit directed vectoring emissions to a first bounded region. Is. The angle between the directed vectoring emissions and the gateway can be, for example, about 45-90 °. In some cases, directed vectoring emissions can be composed of infrared light.

During migration mode, drive system 130 is configured to guide robot 100 towards beacon 150 when robot 100 encounters directed vectoring emissions emitted from vectoring beacon emitter 156. It is possible. Robot 100 can then determine the position of Beacon 150 based on the detection direction of the directed vectoring emissions with respect to the directional receiver 106. Upon determining the position, the robot 100 can align itself in the driving direction with respect to the detected vectoring emissions. For example, robot 100 can move forward along the path of detected vectoring emissions to reach and cross the gateway where Beacon 150 is located.

Robot 100 can remotely activate and stop beacon emissions. For example, the robot 100 can pass start and stop signals using wireless communications such as radio frequency (RF) communications. Robot 100 can remotely activate the directed vectoring emissions of Beacon 150 to stop gateway emissions when it enters migration mode. When the migration mode ends, the robot 100 can remotely stop the directed vectoring emissions of the Beacon 150 to activate the gateway emissions.

In some cases, the Beacon 150 can be configured to exchange schedule information with the Robot 100. For example, the beacon 150 can transmit the time when the robot 100 should enter the cleaning mode, the time when the robot 100 should enter the migration mode, and the like. Schedule information can include details such as start or end times and dates.

The navigation beacon 150 is also capable of transmitting proximity emissions laterally around itself. Robot 100 can avoid cleaning and migration within proximity emissions by performing preconfigured actions such as changing the course when proximity emissions are detected. Melee emissions can be thought of as a "force field" where the robot 100 is not allowed to pass.

The navigation beacon 150 can have a switch to select a confinement mode, a navigation mode, and an off mode. The navigation beacon 150 can have a range selection switch such as a virtual wall. It is possible to indicate with light that the navigation beacon is operating. The navigation beacon 150 can only turn on the motion indicator when permitted by the robot 100 or when instructed to do so. It is also possible to have a low battery warning and a separate low battery lamp.

Different modes of operation can be present in the navigation beacon to reduce power consumption and thus extend battery life. When the robot 100 is not in operation, the navigation beacon 150 can be in low power mode, in which the emitter is off and the navigation beacon periodically monitors the communication link to wake up. Determine if it is needed. When the robot 100 is in operation, it is possible to send signals to a single or multiple navigation beacons to turn on their respective emitters. It is possible to turn on different emitters with different commands. In addition, the navigation beacon can return to low power mode after some elapsed time.

When multiple navigation beas are used in Robot 100, each navigation beacon can have a unique identifier, such as a 16-bit (or higher) identification number contained in memory. .. This identifier can be generated in the factory or by the navigation beacon itself. When the identifier is generated in the factory, it can be stored in the beacon in the non-volatile memory. Robot 100 can use a unique identifier to generate an internal map that can be used to navigate from one bounded area to the other by navigating from one navigation beacon to the other. It can be used for navigating. In some embodiments, the identifier for the navigation beacon 150 can be used to generate an emission-encoded signal that can be detected by the robot 100. If the navigation beacon 150 is unable to transmit the entire identification number, a unique code can be generated as a derivative of the ID or by navigation between the existing navigation beacon and the robot 100.

Robots and navigation beasons can communicate point-to-point or by broadcast transmission. In a point-to-point scheme, the robot must learn all the identifiers of the navigation beacon. It is then possible to transmit a transmission with a data field indicating the destination navigation beacon. Only the navigation beacon and the navigation beacon respond accordingly. This has the advantage of extending battery life as only certain navigation beas react and leave other navigation beas off. User interactions can be used to train the robot to see the navigation beacon.

In an alternative concept, the robot broadcasts a command to enable all navigation beas to all navigation beas. This scheme works without user interaction. In these navigation beacons, another technique reduces the transmission power of the robot in order to increase battery consumption and mitigate the effects of shortening battery life.

FIG. 13 is a diagram illustrating an embodiment of a simplified navigation beacon 152 that can be used with an autonomous mobile robot navigation system. In this embodiment, the beacon has a beacon emitter assembly 152 that includes a gateway emitter 154 and an omnidirectional emitter 160, but no vectoring emitter. Although it is possible to demarcate regions using such beacons, it does not necessarily support the migration feature.

14-17 are various diagrams of an exemplary base station 180 that can be included within an autonomous mobile robot system. The base station 180 can be located within the bounded area provided by the system. The base station 180 has a base 182 with a defined docking direction and a robot charger 184. The robot 100 can approach the base 182 from the docking direction in which the robot 100 is docked with the base 182. The robot charger 184 is housed in a base 182 and can be used to charge the robot 100 when properly docked.

The robot charger 184 can start operating when the robot 100 is docked with the base 182 and then activated by the detection circuit. Battery conditions can be used to control whether charge / discharge mode, precharge trickle mode, or postcharge trickle mode was used to charge the battery.

18-21 are perspective views, top views, rear views, and side views of an exemplary emitter assembly 186 used within the base station 180. Emitter assembly 186 has a lower portion housing 188 that holds an omnidirectional beam emitter 190 and two field emitters 192 and 194 for navigation.

The omnidirectional emitter 190 can be mounted on the housing 188 and configured to project a proximity beam laterally around the docking station through the use of the upper portion housing 196. The upper portion housing 196 can be configured such that emissions from the omnidirectional emitter 190 form a desired pattern, such as a circle, on the outside of the base 182.

The navigation field emitters 192 and 194 can be attached to the lower portion housing 188, each configured to radiate a field of laterally bounded and overlapping signal beam emissions. One of the radiated fields can be aligned with the docking direction to define a lateral field edge that overlaps the other radiated fields. The two navigation field emitters of the base station can be arranged so that their overlapping angle is about 6 ° and the angle of the opening of each emitter can be 20 to 30 °. Protrusions 198 on the lower portion housing 188 can be used to shape emissions from emitters 192 and 194 to achieve the pattern described above. In one embodiment, the navigation field emitter 192 projects a beam at about 12 ° from the normal direction of the emitter assembly 186, which extends from about -5 ° to about 35 °, about 40 °. Other navigation field emitters 194 feature an L-shaped baffle or mask 198 that borders near the middle of the LED and projects a beam at approximately 12 ° from the normal of the emitter assembly 186, which is 0 ~ Spreads to about -35 °. In some embodiments, the emission field can be configured with infrared light (IR). In such cases, the field emitters 190, 192, and 194 can be configured with infrared light emitting diodes (LEDs).

FIG. 22 is a schematic view of the electronic component 101 of the robot 100. The electronic component 101 includes a controller 103 that communicates with an omnidirectional receiver 104, a directional receiver 106, a wall proximity sensor 1072, and a bumper microcontroller 107A that controls a bumper switch 1074. Controller 103 monitors all other sensor inputs, including the cliff sensor 1076.

23A and 23B are diagrams showing an embodiment in which a method of navigating an autonomous coverage robot between bounded regions is applied using a navigation beacon. This method can be applied to a system such as the robot navigation system of the above-described embodiment. The following description relates to a particular software architecture and set of algorithms for navigation beacons disclosed as physical infrastructure, but very similar embodiments of robots, controls, and software architectures, and / or a number of specific. The algorithm can be used with different physical infrastructures.

Therefore, the mode and method by which the robot processes the target and navigation sequence is relatively independent of the actual physical form of the infrastructure, and in the description of such mode and method, the "navigation beacon" is used. , Other physical types of active and passive landmarks (eg, bar codes, back-reflective tags, lighting or IR spots projected onto ceilings or walls, directional RF or visible light, and identification within the environment. Including other possible features), the "beam" contains other physical types of oriented vectors (which can be followed, calculated or held in memory, estimated direction vectors, and). Includes the direction of the line of sight, especially with respect to the observed object). In the examples shown, the wall 201, the first navigation beacon 202, and the second navigation beacon 203 (also referred to herein as "lighthouses") are the first bounded area 204, the first. It defines two bounded regions 206 and a third bounded region 207, which are generally home rooms. The navigation beacon 202 intersects the gateway 210, which is generally a doorway, but can have openings of a few feet to 10 or 12 feet, and also arbitrarily divides the room without a specific opening. It can be configured to transmit gateway marking emissions 208 (also referred to herein as "virtual gateway" or "gateway beam"). The gateway marking emission 208 can be an IR signal and can be detected by, for example, the coverage robot 212.

The robot 212 can be placed in the bounded area 204. The robot 212 can autonomously cross the bounded area 204 in the cleaning mode. When the robot 212 encounters the gateway marking emission 208 at the gateway 210, the robot 212 remains in the bounded region 204 without moving to the adjacent bounded region 206. Upon exiting the cleaning or coverage mode within the first region, the robot 212 autonomously (ie, responds to internal trigger conditions such as time out, response distance, response ratio, etc., as described herein. In this mode, the robot continues cleaning, but monitors gateway 210, detects gateway 210, heads for gateway 210, and passes through gateway 210 (passing beacon 202). Enter the adjacent bounded area 206.

In some embodiments, the robot 212 is capable of autonomously entering cleaning mode after the migration to bounded area 206 is complete. During migration mode, the robot 202 can be unresponsive to gateway emissions 208 (but it is possible for the robot to record that it has crossed gateway emissions 208). In some cases, the robot 212 may be able to remotely stop the gateway marking emissions 208 when entering migration mode (ie, by guided, relayed RF communication, line-of-sight, or reflection of optical signaling). is there. The robot 212 can also activate the gateway marking emissions 208 after exiting the migration mode.

In addition, the navigation beacon 202 is capable of laterally transmitting proximity emissions or fields 214 around itself (periphery) to establish exclusion zones or "force fields". Robot 212 can detect this proximity field during cleaning or migration mode and use that detection to attempt to avoid being moved within the exclusion zone around Beacon 202. The main purpose of the exclusion zone is to prevent the robot 202 from hitting and moving the beacon 202 itself.

The navigation beacon 202 is a first directed vectoring emission 216 (in the present specification, "directed emission", "directed beam", "directed beam", by a beacon 202 arranged within the gateway 210. A start beam (also referred to as a "tractor beam") can be configured to be transmitted to the bounded region 204 and a second oriented vectoring emission 217 to be transmitted to the bounded region 206. For example, the robot 212 can be driven towards the beacon 202 when it detects the directed emissions 216 during migration mode.

During the migration mode, the robot 212 is located at the front of the robot 212 (ie, the side surface of the robot 212 aligned forward in the driving direction of the robot), the directional receiver 218 (in the present specification, for example, parallel). It is possible to detect directed emissions 216 by detecting directed emissions 216 by means of a "binocular sensor") that includes a pair of detectors in the tube. Alternatively, the robot 212 is capable of detecting emissions 216 directed by the omnidirectional receiver 222 prior to detection by the directional receiver 218. The term "omnidirectional receiver" as used herein includes omnidirectional and multidirectional receivers, including, for example, two or more sensors directed around the perimeter of the robot (such as at a compass point). Robot 212 then uses the signal detected by omnidirectional receiver 222 to move to detect emissions 216 directed by directional receiver 218 (in-place turning or forward or backward direction). It is possible to turn to). The robot 212 is capable of aligning its direction of travel with the path defined by the edge 220 of the directed emissions 216. The directional and omnidirectional receivers 218 and 222 can have similar configurations and functions as those described above.

In some cases, the robot 212 moves toward the beacon 202 along the emission path (ie, the two beams of the binocular sensor 106) until it detects the beacon periphery 224 (ie, the edge of the proximity field emission 214). Servo using detectors 114 and 116) is possible. The robot 212 can move to the bounded region 206 along the beam periphery 224 via the gateway 210. The angle between the gateway emission 208 and the directed emission 216 can be about 45-90 ° in some cases and about 60 ° depending on the situation. The gateway emissions 208 are sized based on the robot's diameter and diverge approximately to the width of the robot at the proximity field edge or detection range (eg, so that the robot does not intersect at that position). This can be a width of 10 ° or more, but is determined by the diameter of the robot. The angle between the center of the gateway emission 208 and the center of each oriented emission 216 is about 45 ° or about 60 ° in the two different embodiments of each oriented emission 216. The beam is a divergent beam diffused by a slot mask near the emitter, at about 20 ° to 30 ° (eg, 25 °) or 30 ° to 50 ° (eg, 40 °) in two different embodiments. is there. In some cases, the directed emissions 216 and 217 can be composed of infrared light.

Robot 212 is capable of remotely activating and / or stopping emissions from Beacon 202, such as directed emissions 216 or gateway emissions 208. Various methods can be used to transmit signals between the robot 212 and the navigation beacons 202 and 203.

In particular, the same inexpensive common IR-LED emitter can be used for each of the emitters of Beacon 202 and 203, in addition to recharging for robot 212 or other docks 240. It is effective. The same sensor on the robot 212 can detect all emitters, and different sensors on the robot 212 (eg, other directions, and parallel directional sensors) have different purposes (eg, tracking as described above, The same emitter can be detected for homing and stopping). To distinguish between different emitters, it is possible to encode each emitter (eg, modulate it with a different serial code). This also helps to avoid confusion due to household and other remote controls, as well as the IR components of sunlight and other environmental sources. However, as shown in FIG. 23, the robot 212 may encounter situations where it travels through the emission paths of multiple different emitters.

In this situation, the sensor or software structure receives multiple signals at the same time, especially when using an optical multidirectional receiver or a common optical omnidirectional receiver (and in other cases, eg RF). , One emitter is configured to be distinguishable from the other. If different frequencies are available, they can be used. According to one method, emitters are used in turn to synchronize by communication with a robot or between emitters. However, this method is complicated and goes into failure mode when synchronization fails. The respective beacons 202 and 203, as well as the dock 240, are self-regulating, and it is convenient to timely shift their own signals at intervals between their respective emissions. For example, each transmission emission is distinguished by time division multiplexing (ie, multiplexing as a beam origin, with a random offset over a period of time (eg, it can be different for each beam source). It is possible (think of it as a system) that includes a beacon with two or more beams or multiple beacons. The interval or random offset can be changed from time to time (eg, at random intervals) or by communication with the robot (eg, by RF communication when the robot detects an interference condition). During manufacturing or changing offsets, for example, offsets can be selected from a set of offsets that are unlikely to interfere in harmony or share common factors. In this way, the robot can encounter multiple emission paths and further identify their identities. In other cases, each transmission emission can be distinguished by wavelength differences in light, infrared modulation, and wavelength filters / windows on emitters and receivers.

Robot 212 can use RF communication to signal Beacon 202 to activate remotely directed emissions 216 to stop gateway emissions 208 when migration mode begins. .. In another embodiment, the robot 212 is capable of stopping the remotely directed vectoring emissions 216 and activating the gateway emissions 208 when the migration mode ends.

In some cases, the robot 212 may activate the gateway emission 226, which separates the bounded areas 206 and 207, to initiate the cleaning mode. Similar to those described above, the robot 212 can be kept out of bounded area 206 when encountering gateway emissions 208 and 226. Upon completing the cleaning of the bounded area 206, the robot 212 enters migration mode and navigates through the gateway 228 by the directed emissions 230 to the peripheral 234 of the proximity emissions 232 to the bounded area. It is possible to enter 207. Upon entering bounded area 207, robot 212 is capable of re-entering cleaning or working mode. For example, after a set amount of time or after encountering a preset number of gateway emissions 208, the robot 212 can migrate from bounded domain 204 to bounded domain 206.

Alternatively, it is possible to configure a virtual wall emitter (gateway) to stop itself independently on a schedule so that the robot can cross the gateway, for example (first). Acts as a virtual wall during the first interval (as limited to the current room, or), then temporarily or intermittently stops, and then (limited to the second or next room) As) during the second interval, acting as a virtual wall. The technique can also be bidirectional, for example, a robot can use RF communication with one or more beasons to stop a virtual wall or bypass / cross a virtual wall. Is.

In all cases herein, the robotic behavioral system is configured such that failure events, including cliff detection events, have a higher priority than other room-to-room navigation or cleaning. Please note. So, for example, when the robot detects a cliff in the middle of a navigation or other sequence of actions, the robot still cliffs (stopping parts of the current sequence and resetting the state of the sequence). Can be avoided.

FIG. 23 shows a series of rooms similar to FIG. 22, but to show the situation when the robot encounters multiple beams from two beacon emitters and dock 240 (with a base station). ) It is a constructed diagram. By using the method described above to avoid confusion, the robot 212 still navigates from room to room, or the robot 212 autonomously enters docking mode and bases within bounded area 206. It is possible to move towards station 240 and dock to the station when the migration mode in the second area ends.

The base station 240 can include a base, a robot charger, an omnidirectional beam emitter, and two navigation field emitters, which may resemble the base station 180 described above. Robot 212 detects one of the lateral field edges 242 or 244 of overlapping fields 246 and 248 aligned with the docking direction and advances along it until docked with station 240. It is possible to move towards the base station 240.

Robot 212 now detects emissions from base station 240 by means of an omnidirectional receiver 222 on the robot to detect lateral field edges (eg, 247) outside of at least one field emission 246 or 248. It is possible to move it. Robot 212 can then advance along the outer lateral field edges 247 or 249 to the aligned lateral field edges 242 or 244 of the overlapping fields. Upon detecting the aligned lateral field edge 242 or 244, the robot 212 advances along the aligned lateral field edge 242 or 244 until docked with the base station 240.

FIG. 24A is a diagram showing the behavioral software architecture within the controller 250, showing a simplified form of goal-oriented or sequenced behavior. FIG. 24B is a diagram showing the same software architecture, but details some of the elements of FIG. 24A with respect to goal-oriented or sequenced behavior.

The robot embodiments described herein can use little or no behavior-based control, but behavior-based control controls the robot robustly (ie, does not stack or fail) and safely. It is effective when doing. Therefore, Figures 24A and 24B show software architectures that can improve control of mobile robots. Rather than a simple reaction architecture, Figures 24A and 24B introduce planning elements that are partially responsive (and therefore more robust), goal-oriented, sequenced reaction behaviors.

In responsive and sequenced behaviors, the implicit success of an associated behavior in a sequence is (especially) the initiation behavior in the sequence, where some sequences can have more than one initiation behavior or branch. Permit conditions for each of the other actions. Implicit success is held in the form of flags, state machines, etc., as states that indicate the conditions under which continuous actions are permitted. This state can be set or changed by the action itself, by the arbiter process, or by the planning process upon completion.

With reference to FIG. 24A, the robot 302 uses a control and software architecture with multiple actions performed by the arbiter 252 within the controller 250. The action is input to the arbiter 252 in response to a sensor event or a sequence state event from the planner 253. In one embodiment, all actions have relative priorities fixed to each other. Arbiter 252 (in this case) recognizes the permit conditions and selects the action that has all of the permit conditions and has the highest priority among those that meet the permit conditions. The figures shown in FIGS. 24A and 24B do not necessarily reflect the (fixed) priority order of robot 302. To lower priority, actions are generally escape and / or avoidance actions (such as avoiding cliffs or escaping from corners), working actions (eg, wall following, bouncing, or linearly driven). ), And sequenced reaction behaviors for this application. This behavior includes different escape behaviors 255, 255a, 255n (eg, U.S. Pat. No. 6,809, Includes a temporary fire-and-foget movement of "shock" that suppresses corner escape, anti-canyoning, stacking, and some evasive behavior, as disclosed in No. 490, see its contents. (Incorporated in its entirety by), Cliff Avoidance 256, Virtual Wall Avoidance 258 (Virtual Wall can be a beacon with a gateway beam), Other Avoidance 258n as a class (eg, for a person by pyrometer) Avoidance, in general, detection of objects that are dangerous to or potentially dangerous to Robot 302 and a class of movement away from it), Spot Coverage 264a (spiral, or pyrophyllometer) Cover with closed pattern like patch), align 264b (in-situ swivel, use side proximity sensor to align with obstacles encountered in front, while following obstacles (eg corners) (Inside)), Follow 264c (indicating approximately parallel follow, bump follow, or both along an obstacle using a lateral proximity sensor or a bumper that extends to the side of the robot), "Bound Includes bump response 264d (action that occurs after the robot hits an object), and drive 266. The movement of robot 302, if any, occurs while the movement is arbitrated. If there are more than one action in the arbiter 252, the action with the higher priority will be performed if any corresponding required condition is met. For example, the cliff avoidance action 256 is not executed unless the cliff is detected by the cliff detection sensor, but the execution of the cliff avoidance action 256 always takes precedence over the execution of other actions that also satisfy the permission condition. If there is, an action with a higher priority is performed if any corresponding required condition is met. For example, the cliff avoidance action 256 is not executed unless the cliff is detected by the cliff detection sensor, but the execution of the cliff avoidance action 256 always takes precedence over the execution of other actions that also satisfy the permission condition. If there is, an action with a higher priority is performed if any corresponding required condition is met. For example, the cliff avoidance action 256 is not executed unless the cliff is detected by the cliff detection sensor, but the execution of the cliff avoidance action 256 always takes precedence over the execution of other actions that also satisfy the permission condition.

Reactive behaviors have detection of various sensors and phenomena as their permit condition or trigger, but are generally not the (arbitrary) state of the sequence. As shown in Figure 24A, these are forward proximity detection (s), forward bump detection (s), cliff sensors (s), virtual wall signal detection (which can be considered as coverage triggers instead). Includes sensors for obstacle avoidance and detection, such as. These types of sensors are monitored and tuned by filters, tunes 292c, and their drivers 292b to help allow permission conditions and actions to take place as expected and all available information (eg, 1 bit). Conversion to a "true / false" signal, recording of possible impact or incident angles based on intensity or time differences from a group of sensors, or recorded data on history, mean, frequency, or change information) Can be generated.

The actual physical sensor can be architecturally represented by a "virtual" sensor synthesized from tuning and drivers. Additional "virtual" sensors are unique to Robot 302 (by monitoring the detectable or interpreted physics and overcurrent of the motor, lack of odometry reading from the wheel encoder or counter). Synthesized from accepting or stacking states, battery charge by coulometric analysis, and proprioceptive or interpreted robot 302, such as other virtual sensors "Virtual N", and other virtual sensors "Virtual N". To.

In addition, reaction behavior can act on the basis of permit conditions that represent the detected phenomenon to be searched for or followed. Beam or radio (RF, acoustic) signals can be detected regardless of direction (and in some cases direction). A directional remote beam or marker (recognized by a vision landmark, bar code, back-reflective, distinctive, reference, or natural) can allow homing or associated movements in the direction. Regardless, the robot 302 can still move to servo with respect to the presence, absence, and / or relative intensity of the detected signal. The reflection of the beam from the robot 302, edge, or line can be detected as well, and the action performed by servoing on such a signal (such as following an obstacle by the robot 302) is continued. Can be done. The signal of the strip or artifact can be recovered by the robot monitoring the strip or object that has been picked up or traversed, and that signal can be a permit condition for reaction behaviors that control the spot coverage pattern. it can. A general class "search / servo N" signal is therefore also a reaction action that allows a condition or trigger.

Robot 302 maintains "parallel" processes, parallel processes that are not generally considered reactive behaviors. The scheduler 292d may need to allocate processor time in a collaborative or other multitasking manner to most other processes, including, for example, arbiters and actions. If more threads are available, it is possible to have fewer processes managed by the scheduler 292d. As mentioned above, the filter and adjustment 292c, as well as the driver 292b, can interpret and convert the unprocessed signal. These processes are not considered reactive behavior and do not extend any direct control over the motor drive and other actuators. In addition, in this embodiment, the brush motor controller 292a controls the main and side brushes, which can be alternately controlled by dedicated brush behavior and brush control arbiters.

With reference to FIG. 24B (shown similar to FIG. 24A and highlighted / deemphasized for clarity), the special parallel process is Sequence Planner 253. The sequence planner 253 can monitor and / or maintain the state to control the sequenced reaction behavior. This action can be performed, for example, between setting permit conditions (captured by Arbiter 252) and / or resetting the state in the case of a stop condition (eg, timeout, escape from action execution or avoiding it), and / or between states. Monitoring or management of finite state machines that migrate themselves. It should be noted that a controlled state machine requesting a transition between permission condition states or flags, or sequential reaction actions, can be set by the action itself, by the arbiter, or by the sequence planner 253, based on efficiency and environment. I want to. The sequence planner 253 also directly controls the arbiter 252 (or the second arbiter 252A) to select an action (eg, prioritize again), or allow conditions, state variables, depending on the situation. It should also be noted that it is possible to directly control the sequenced reaction behavior without setting a state machine.

It is recognized that preserving state potentially reduces the robustness of the component's reaction control architecture, but the sequence of actions of interest may require maintenance of some state. As described herein, when the sequenced reaction behavior is controlled and executed using a request state that can be robustly reset or deleted without disrupting key functions. The loss of robustness can be reduced.

An exemplary sequenced response behavior setting is used by robot 302 to move between rooms or navigate to a distal room. As shown in FIG. 24B, the homing beam alignment action 294a includes a permission condition for the alignment start (eg, homing request) to be activated and the beam to be returned detected by the multidirectional sensor. Homing request authorization conditions can be set and retained in different ways, as described below. Homing action 294b of the homing beam is performed when the alignment is successful. Note that the homing condition of the homing beam is a homing request (flag or state) as a permission condition because alignment can occur unexpectedly (the robot can already align when the homing request is activated). ), And using directional sensors (binoculars) that are substantially and inevitably directed (or filtered to be equivalent to, for example, such detection) to the beam identified as the beam to return. Also has detection. Therefore, the two sequenced reaction actions are separate entry positions for the sequence and share the permission conditions for a common reaction sequence request. If the sensor-based authorization conditions fail, these actions are aborted (internally or by the planner 253) and the sequence request remains active (or can be reset). In this case, the sequence can be restarted the next time any sensor-based authorization condition arises. Depending on the situation, random components (returning to random coverage, while disabling the allowance conditions for sequence or homing requests, random distance or time) can lead to systematic glitches or looping. To reduce, it can occur with the discontinuation of sequential reaction behavior.

After the homing action 294b of the homing beam is completed, the sequence is assumed for a condition that allows the robot to start field tracking around the beacon so that it is properly positioned and aligned with the field and transitions to a new room. It is in a state. The (new) alignment start permission conditions (eg, field alignment state or request) can be set by completing a sequenced reaction action or by the planner via a condition flag or state machine. Proximity field detection by the multidirectional receiver is under the same (OR) conditions as the field alignment request so that the field can be the first detectable signal encountered at the start of the homing request. If there is, and the homing request is active when the field is encountered, this combination also initiates field alignment action 294c. If there is no stop condition, the field alignment action 294c is executed. After the field alignment action 294c is completed, the sequence (of the sequenced reaction action) is in a state where the following assumptions are made regarding the state in which the robot is properly oriented and begins field tracking. Allowed conditions for (new) follow-up initiation (eg, field-following states or requests) can be set by completing a sequenced reaction action or by the planner via a condition flag or state machine. Similarly, the initiation action 294e can have a state-based request rather than a sensor detection as a primary permit condition. Each of these actions 294a, 294b, 294c, 294d, 294e will be described in detail below.

In two of these cases, depending on the situation, the state to enter the entire sequence (homing request) is retained, and further states to start the next part of the sequence (field follow request, start request). ) Is set, and both conditions need to initiate the next sequenced reaction action. This allows the sequence to be resumed immediately after aborting (by preserving the entire sequence request state). Alternatively, the entry request for the entire sequence is reset and only the entry conditions for the next sequenced reaction action (field alignment request or field follow request) are set. This causes the entire sequence to be restarted separately (for example, due to new or continuously occurring conditions that cause the planner to set a sequence request) after the abort, basically by clearing all states. It is possible to improve the strength (although not necessarily the strength that specifies the target).

In two cases, no sensor-based condition is a permit condition for the next sequenced reaction action in the sequence, but the permit condition is a sequence-based or state-based request (eg, field-following request, start request). ). Therefore, a sequence is a sequenced reaction behavior that includes only a sequence state or progression as a permit condition for the next sequenced reaction behavior (field-following behavior, start behavior, etc.), and as a permit condition (beam alignment or tractor homing). It is possible to include sequenced reaction behaviors, including behaviors), sensor-based conditions and sequenced states, progressions, or requests.

In some cases, the state for entering the entire sequence (homing request) is retained, and either the sensor-based condition or the sequence-based request is due to the next sequenced reaction behavior (field alignment) in the sequence. It is a permission condition of. For example, homing request AND (field alignment request OR field detection) can be interpreted by the arbiter to perform field alignment in the absence of higher priority reaction behavior. Thus, a sequence may include a sequenced reaction behavior, including any combination of sequence states / requests, as a permit condition for the next sequenced reaction behavior, or a sensor-based condition and sequence request /. It is possible to include both states.

Note that entry and intermediate sequence branching is possible by allowing different actions based on different combinations of sensor-based conditions and sequence states. For example, two possible entry positions in a sequence are sequenced differently when they are allowed by detecting different sensor events (multidirectional or directional) at the same time as the common entry position permission condition (homing request). Branching within a sequence for a reaction action is allowed by detecting different sensor events at the same time as the permission conditions of a common intermediate sequence (eg, candidate actions allowed by a unique combination of sensor events and sequence states, respectively). The permit. Depending on the situation, if each of the different sequences is held in a different state, the combination of states in the different intermediate sequences allows nesting of the sequence (each of the candidate behaviors allowed by the unique combination).

The sequence planner 253 holds a state switch or finite state machine 243a in navigation mode, depending on the situation, and / or such states or requests are held separately (eg, follow start, follow complete).

As shown in FIG. 24B, the planner 253 is capable of controlling different goal-oriented sequences 1-N, each having two or more sequenced reaction behaviors within it. Actions within it can have target states, entry requests, or other states as initial conditions, and Arbiter 252 identifies candidate actions available for execution as required renderings. A sequence request may be required for the initial action of the sequence. Subsequent actions may require one of the following sequenced reaction action requests (or, in addition, a sequence request).

Sequenced behaviors depend at least on the same arbiter 252 as reaction behaviors, and / or respond to sequenced states as a permit condition similar to reaction behaviors that respond to real-time sensor events, and / or at least escape. And because it may be prioritized lower than avoidance behavior, it is a "reaction" even though it depends on the sequence state. In addition, it is possible to completely reset the state and restart the reaction process. Even if the permission condition of the sequence or the next sequenced reaction action is satisfied, the reaction action that also has a satisfactory permission condition and maintains continuous autonomous movement is executed instead. Reactive docking behaviors generally allow the robot to return and charge for continued autonomous movements, and are given higher priority than navigation-related sequenced reactive behaviors in particular. Reaction coverage behaviors can, in particular, be prioritized below the sequenced reaction behaviors of the navigation. Some parts of the docking behaviors described herein are substituted as sequenced reaction behaviors. Alternatively, other sequences are possible. Beacon-based navigation is an embodiment of a combination of sequenced and unsequenced reaction behaviors, as described herein.

As shown in FIG. 24B, the second arbiter 252a is capable of managing a series of sequenced reaction behaviors (eg, in combination with the sequence planner 253). Alternatively, the master arbiter (not shown) can manage the first arbiter of the reaction behavior (eg, 252) and the subsequent arbiters of different sequenced reaction behaviors (eg, 252a).

As shown in Figure 24B, especially for navigation applications, the sequence planner 253 can include a monitor, hold a position state in a topological form, or include its equivalent for the environment in which the robot operates. is there. This can be a topological map with nodes and links by using beacons, where the nodes are rooms (thus the links are gateways or beacons) or gateways. (Thus, the link becomes a side of the gateway). The primary use of positional states is to provide a homing beam or homing reference identity to the arbiter or to a sequenced reaction behavior. When provided to an arbiter, the arbiter can only recognize sensor-based permission conditions for behavior (eg, beam detection) if the beam has the proper identity (symbolization) for the position state. is there. Alternatively, the action itself examines the identity or encoding of the received authorization conditions (eg, beam detection). A stop condition can occur if the beacon is reconfigured, etc., and this stop condition can cause a blacklisting of the beacon or criteria, or a map reset (requires new mapping). it can. In either case, robot 302 continues to provide coverage and other remaining reaction behaviors in the local area. The Planner 253 uses the position state to traverse the distance to the remote or distal room (to return from the dock for charging, to go directly from the dock to the remote or distal room). To perform a sequence of sequenced reaction actions in succession, and to provide a guide reference or beam identity in the proper order for the navigation direction based on the direction to move on the topological map. For navigation in different directions (as long as the position is retained)

One candidate routine held / executed by Planner 253 is shown in Figure 25A. The elements of this routine are not always in order (for example, different states may deviate from the indicated order). Planner routines monitor goal-oriented or sequence start conditions (502). As mentioned above, this can be a recording variable of time, distance, cleanliness, measurement, history of inter-obstacle interference, or other criteria and is planned when the conditions are met. Indicates that a separate sequence (such as room-to-room navigation) should be performed. The condition can be a state, a finite state machine flag or a state. If such a condition exists (504), the planner checks the state of the robot according to the indicated sequence to see if the planner holds such a sequence state (506). For example, the planner 253 can hold or monitor the current state of a state map, or finite state machine that defines the state of the beacon crossing sequence, as shown in Figure 25B (here, "for start". The states of "Alignment", "Homing", "Alignment after homing", and "Following field" are available), and alternative state maps or finite state machines can be exemplified. Planner 253, in addition or alternative, sets conditions for permitting action for entry into the sequence, such as homing requests as described herein (508), or the entire sequence. It is possible to set the permission conditions for entry actions and / or the permission conditions for the following sequenced reaction actions. The planner 253 additionally, for example, based on the identification state of the map (eg, as shown in FIG. 25C, of lighthouse / beacon X, Y, Z beams A or B- "green" or "red"-. Based on the state of the map, which defines which one is connected to rooms 1, 2, 3 and the dock, and based on the next target room set by Planner 253. It is possible to set guide variables for the behavior or arbiter to recognize (eg, the beam to follow, or the identity of the reference to go back or forward) (510). Planner 253 additionally sets the conditions for discontinuing the reaction sequence sequenced in progress (512, eg, timeout, interruption due to reaction avoidance or escape behavior), and / or (the behavior internally sets these prior to discontinuation). It is possible to monitor the state of the map at the time of such a stop (although it can be reset).

On controller 250 after the robot 302 detects a specified cleaning period, or after the robot 302 detects a predetermined number of collisions, or after the robot detects certain conditions required for a series of sequenced actions. The sequence planner 253 running in is determined to exit cleaning mode and enter migration mode. The robot can also enter migration mode directly from the dock to proceed there to begin cleaning the distal or remote room. Migration mode performs the task of moving from one room to an adjacent room. As mentioned above, the migration mode is conveniently not a monolithic process, but is a series of sequenced reaction actions initiated by the initial action, which is also referred to as a migration mode request (also referred to herein as a "homing request"). ) Permission conditions. Planner 253 sets permissions, states, or flags for "new room" / homing requests. The remaining permit conditions are sensor detection of the guided tractor beam (depending on the situation, limited to one beam identified based on position state 253b, starting with the lighthouse / beacon to return). Trigger the homing beam to perform alignment action 294a almost immediately if the robot is in the beam, or after the robot encounters the beam while meeting the permit conditions and continuing the coverage action). To be selected by Arbiter 252. Robot 302 performs a homing beam (alignment action 294a) and, at the same time, optionally (when either cliff detector is unable to recognize the floor, i.e., satisfies the permit condition for a higher priority reaction avoidance action). Migrate from one bounded region to an adjacent bounded region by performing Cliff Avoidance 256 (in all cases) and also performing Drive 266 actions (eg, if the abort condition resets the sequence). Do the task. Multiple actions are sensors When triggered by an input or planner 253, the arbiter 252 arbitrates the actions in order of priority. That is, the arbiter can arbitrate in different sequences of reaction behaviors, or the planner can only start one sequence. The beacon intersecting sequence 262 has five sequenced reaction actions, starting with a homing beam (alignment action 294a), which actions are based on controller inputs as described herein. And are executed sequentially. If the robot 302 receives a signal input (such as a bump sensor input) that requires the execution of a higher priority action, the robot 302 will perform the higher priority action and then the required permission. Resume the previous sequence of sequenced reaction behaviors when the condition exists.

31A-G are bird's-eye views showing a behavior-based autonomous coverage robot 302 when moving from one area to another using the navigation beacon 304. Figures 26-30 show beacons intersecting a sequenced reaction action 262, including beam alignment 294a, beam homing 294b, field tracking 294c, field alignment 294d, and beacon departure 294e actions.

Sequence 262 intersecting the beacon begins by entering beam homing 294b or beam alignment action 294a based on the robot's position within the first bounded region with respect to beacon 304. Generally, the robot 302 initiates maneuvering and its omnidirectional receiver 104 detects a first oriented beam 306 from the navigation beacon 304 to the first region (eg, for example. 360 ° swivel or drive in a circle).

FIG. 26 is a diagram showing the beam alignment behavior 294a. As shown, the arbiter 252 is provided by the multidirectional or omnidirectional sensor 322 of the robot 302, at least until the homing request state or request is indicated, and the directional or tractor beam 306 as a permit condition or trigger for action. Do not allow beam alignment action 294a as a candidate (ie, prioritize other actions that meet the permit conditions) until received. No high-priority behavior (eg, reaction behavior) has a satisfactory permit condition to permit execution. This tractor or oriented beam 306 is expected at the current position of robot 302 on the state map (for example, by corresponding only to beams with a code corresponding to the identity or current map state). It is possible to limit to. When these conditions are met, the arbiter 252 permits the beam alignment action 294a to be performed. If either condition is no longer met, it is possible to stop the action. As mentioned in FIG. 26, discontinuation conditions (including, for example, permission conditions for missing actions, timeouts, or replacements with higher priority actions) can be monitored by the planner 253 and of the sequence. It is also possible to hold the state reflecting the progress and the position of the robot 302. The beam alignment action 294a basically rotates or turns the robot 302 on the spot (522) and departs in a random direction. Possible beam detections can also be retained, for example, based on past beam detections monitored by the current process, in which case the swivel direction is expected to be shorter. Is possible. When the directional receiver 318 receives the directed beam 306 (524), it terminates its action. Robot 302 encounters the appropriate beam unless the homing request is active and attempts to align and return when either is a valid entry position for the sequence. Since it continues, there is no need to set the state as a request for beam homing action 294b that may follow. The stage after the sequence is an intermediate sequence, and it is possible to use such a state. Therefore, beam alignment behavior 294a is an embodiment of selection and does not set the state to maintain a higher level of reaction control and possible robustness.

Figure 31A shows the preparation for beam alignment action 294a when robot 302 encounters the first oriented beam 306 emitted by the navigation beacon 304 when the homing request from planner 293 is active. It is a figure which shows the execution. Upon detecting the first oriented beam 306, the robot 302 initiates the beam alignment action 294a. During beam alignment 294a, the drive system 130 moves the robot 302 (eg, by in-situ turning, eg, by departing in a random direction), and the beam emission 306 first oriented by the directional receiver 318. Is detected and aligned with the driving direction of the robot. By using the signals from the two components receivers 114 and 116 of the directional receiver 318, the robot 302 aligns the drive direction with the oriented emissions 306 of the beacon 304. FIG. 31B is a diagram showing that the robot 302 detects the beam 306 by the omnidirectional receiver 322 and turns so as to face the beacon 304.

After the robot 302 aligns the drive direction with the directed emission 306 of the beacon 304 by using the directional receiver 106, the robot can initiate the beam homing action 294b.

FIG. 27 is a diagram showing a tractor or directed beam homing behavior 294b. As shown, the arbiter 252 is at least until the homing request state or request is activated and the directional or tractor beam 306 is received by the directional sensor 318 of the robot 322 as a permit condition or trigger for action (most). Do not allow execution (of high priority candidates). This tractor or oriented beam 306 can also be limited to what is expected at the current position of the robot 302 on the state map. If these conditions are met, the arbiter 252 can allow the beam alignment action 294b to be performed. Action 294b can be discontinued if either condition is no longer met. As described in FIG. 27, the discontinuation condition and the obtained state can be managed by the planner 253. The tractor or directed beam homing action 294b follows the directed beam emissions (532). One exemplary method of tracking uses alternating curves. The parallel sensors 114 and 116 of the directional binocular sensor 318 each reflect the presence of a beam based on the intensity of the signal (eg, independently thresholded or mutually thresholded). Adjusted to go through a false value. True-false, true-true, and false-true beam detection is possible as the robot 302 turns towards and away from beam 306 (adjustment allows true-false and false-possibility). It can be truly limited). "Bang-bang" servoing can be used to change the turning direction as the robot 302 moves forward. Each swirl is part of an arc with a reduced radius, an inward spiral. Alternatively, it is possible to retain analog values and use more complex feedback controls. Follow-up (532) means that the robot moves in the right direction and moves straight forward. Alternatively, the frequency and dispersion tracking described herein can be used with respect to docking to determine. When the omnidirectional / omnidirectional receiver 322 detects the directed beam 306 (534), the action is ready to end. The state is set to continue the sequence of sequenced reaction behaviors, i.e., the sequence of field alignment requests. Action 294b can set this state itself, for example, as a flag in memory (536). Alternatively, the planner 253 can monitor that action 294a is complete and such a flag or state is set in a finite state map or machine.

The behavior sequence, as stated, is that if the directional receiver 318 has already detected the directed beam emission 306 and is aligned with the already oriented beam emission 306, then the robot. Allows 302 to omit beam alignment action 294a after initiating beacon crossing sequence 262. The discontinuation condition can be avoided so that it can be restarted later. As described herein, a combination of reaction-behavior-based control and sequenced reaction-behavior control is beneficial. However, some embodiments described herein do not necessarily exclude other reactive or non-reactive techniques. During the beam homing action 294b, the robot 302 moves towards the navigation beacon 304, while responding to the signal of the directional receiver 318 by a directed beam (eg, servoing). Hold that alignment. FIG. 31C is a diagram showing a robot 302 approaching the navigation beacon 304. As the robot 302 follows the edge 320 of the beam 306, the robot 302 adjusts its course as indicated by the arrow 308. As shown in Figure 31C, Robot 302 stops moving towards Beacon 304 when it encounters a proximity field emission (or "force field") projected laterally from Beacon 304 in all directions. To do.

Another behavior that is preferred as part of the sequenced reaction behavior is that the directional receiver 318 is more remote and more sensitive (eg, by collimation and confinement) than the omnidirectional receiver 322. When selected, the directional receiver 318 is capable of receiving force field emissions 310, even though the omnidirectional sensor 322 does not receive. In this case, the directional sensor 318 can be used to alternatively return onto the force field emission 310. In this case, except that another action is provided in the sequence and the sensor-based authorization condition is that the force field 310 is recognized within the directional sensor 318, and that homing is on this emission 310. It is the same as the tractor homing action 294b. As an example of behavioral robustness, this behavior can be arbitrarily added to a sequence of reaction behaviors without altering other behaviors or interrupting other processes, making it more reliable. Based on any empirical basis, it has a higher or lower priority than the existing tractor homing behavior 294a.

Upon detecting the proximity field 310, the robot 302 executes the field alignment action 294c. FIG. 28 is a diagram showing the field alignment behavior 294c. As shown, the arbiter 252 will (mostly) until at least the homing request state or request is activated, one of the field alignment requests is active, or the proximity field 310 is recognized by the omnidirectional sensor 322. Do not allow execution (of high priority candidates) (the latter is field alignment 294c when the action sequence accidentally "meets" the proximity field 310 before the robot 302 encounters the oriented beam 306. Allow to proceed to). The field can also be limited to what is expected at the current position of the robot 302 on the state map. If these conditions are met, the arbiter 252 can be allowed to perform the field alignment action 294c, and the discontinuation condition is managed by the planner 253 as described above. The field alignment action 294c basically rotates or turns the robot on the spot (544). Possible beam directions follow the beam identity (eg, towards the dock 340, following the left beam 350 relative to the dock 340, the robot 302 turns clockwise, and the right beam 360. This means turning counterclockwise), or other conditions, or can be informed. Robot 302 is capable of turning in a direction that is expected to be shorter or at random. The omnidirectional or omnidirectional receiver 322 detects the directed beam 306 (546) and terminates action 294c. Depending on the situation, action 294c does not end until further conditions are met, and the directional beam 306 is not present in the directional receiver 318 (this is the robot 302 rotating across the dock 340). Increase the likelihood of doing so). The field follow-up request is the action itself, as described above.

Thus, as shown in FIG. 31D, when the robot 302 faces the beacon 304, the robot 302 initiates a swivel maneuver (eg, has a drive wheel over the center-to-center distance or has a horomic drive. Based on the beam coding of the first oriented beam 306, in-situ swivel for differentially driven robots, or highly maneuverable robots, substantially in-situ swivel for other configurations). By changing the driving direction of the robot to the right or left towards the other oriented beam 314, the robot 302 can detect the proximity field 310 by the directional receiver 318. The robot 302 stops turning when the directional receiver 106 no longer detects the proximity beam 310.

Robot 302 may encounter proximity field 310 before encountering directed beam 306. In this case, the robot 302 does not perform the beam alignment action 294a or the beam homing action 294b. Instead, the robot 302 recognizes the coding of the proximity field 310 and performs the field alignment action 294d.

When robot 302 no longer detects the proximity field beam 310 by the directional receiver 218, robot 302 performs field-following action 294c.

FIG. 29 is a diagram showing the field following action 294d. As shown, Arbiter 252 does not allow execution (of the highest priority candidate) at least until the homing request state or request is active and the field follow request is active. As mentioned above, this can be simply limited to field-following requests if the sequence is more easily abandoned and remains robust. The fields are limited to what is expected, depending on the situation. When these conditions are met, the arbiter 252 can allow the execution of field-following action 294d, which is dependent on the stop condition. Field-following action 294c follows edge 312 of proximity field emission 310 using alternating curves with reduced radii as described above (554). The omnidirectional sensor 322 can be configured to return a 1-bit truth-false value that reflects the presence of the beam, based on the signal strength. When the omnidirectional / omnidirectional receiver 322 detects the next oriented beam 314, the starting beam (534) (this detection is limited to the expected beam of beacon 304, and / or (Based on the position state of the robot 304), the action is ready to end. Depending on the situation, the directional receiver 318 will prematurely start the beam 314 (due to reflections from highly reflective surfaces, such as nearby white doors or corners of walls, the omnidirectional detector 322 will start beam 314. Action 294d proceeds with further checks until it no longer detects the start beam 314 (which tends to exclude cases of detection). The state is set to continue the sequence of reaction actions, i.e. the sequence of start requests. Action 294c can set this state itself (558), or planner 253 can control it.

As shown in FIG. 31E, during the field-following action 294d, the robot 302 is moved in the resulting arc around the beacon 304, while the edge 312 of the proximity beam 310 via the directional receiver 322. Follow. The robot 302 follows the edge 312 of the force field 310 by monitoring the received signal strength of the force field emission 310 on the robot's emission sensors 318 and 322. One example of moving an arc involves the robot 302 moving forward and backward in a "zigzag" along the edge 312 of the proximity beam 310 (eg, using "bang bang" servoing, or 1-bit true-false detection. Following, and initially arcuate forward motion (eg, spiral) that gradually reduces the turning radius, the robot 302 detects the entry and exit of the proximity beam 310 into and out of the field by means of the omnidirectional receiver 322. .. In another embodiment, the robot 302 has a PID (Proportional-Integral-Derivative:) based on the time in the force field 310 or the timeout of the force field 310. Use the Proportional-Integral-Differential) control system to follow the force field edge 312. The PID control system controls the arc shape of the robot 302. Robot 302 performs reduced speed (eg, less than 50% of coverage speed) maneuvers on both drive wheels 132 and 134 so that both drive wheels 132 and 134 are driven in the forward direction. .. This facilitates transitions and crossing thresholds. Robot 302 passes through an area normally covered by gateway beam 316. As shown in FIG. 31, the omnidirectional receiver 322 detects that the robot 302 enters the second directional beam 314 and the directional receiver 318 no longer detects the directed beam 314. Until then, the robot 302 continues to follow the edge 312 of the proximity beam 310. As mentioned, meeting both criteria helps the robot 302 not detect beam emissions reflected by adjacent objects such as white doors.

FIG. 29 is a diagram showing the start operation 294e. As shown, Arbiter 252 does not allow execution (of the highest priority candidate) until the start request is active and a homing request may also be needed. The start motion 294e assumes its final position by the omnidirectional detector 322 within the start beam 314 and also traverses the start beam 314 approximately and curves away from the field 314 / beacon 304 in the approximate direction of the start beam 314. (For example, if the robot 302 crosses from left to right and curves clockwise, and if the robot 202 crosses in the opposite direction and curves counterclockwise, it faces the dock 340. ). Depending on the situation, the robot 302 attempts to stay within range of the start beam 314 by servoing back to the start beam 314 when the omnidirectional sensor 322 does not detect the start beam 314. Is possible. After a given amount of time (eg, 5 seconds) or distance (eg, 1 m), or contact with an obstacle, the action is ready to end. At this point, the planner 253 is capable of terminating the homing request (and, for example, updating the position state of the topological map and, if retained, clearing or resetting the state of the navigation mode) and reacting. Begins coverage work for the robot in subsequent rooms (or, if the battery is charged, it is possible to initiate another homing request in the direction of the dock, based on the map's position status). The planner 253 also immediately initiates another homing request, or maps the robot 302 to move from room to room, for example, directly into the room or from a distal or remote room. It is also possible to reset the variables of the current homing request based on the position state of.

That is, when the field-following action 294c is completed, as shown in FIG. 31F, the robot 302 can perform the exit beacon action 294e, as shown in FIG. 30, which means that the robot 302 is a beacon. Accompanied by movement away from 304. Alternatively, this can be performed as a swivel and start, and / or guided by a start beam 314 and an omnidirectional receiver 322. For example, FIG. 31G shows that as the robot 302 passes through the edge of the oriented beam 314, the robot 302 turns away from the navigation beacon 304 and faces a second region. Is. If the robot 302 detects a second oriented beam emission 314 by the omnidirectional receiver 322, the robot 302 is moved away from the beacon 304. If the robot 302 does not detect the beam emission 314 second oriented by the omnidirectional receiver 322, the robot 302 avoids arcing back to the gateway, as shown in Figure 31H. Move forward without turning to. Another embodiment of the withdrawal beacon action 294e is that the robot 302 resumes the sequence of actions when the mileage measurement method (or, for example, the planner 253 aligns the beam with the beam or map field of the next beacon). Of the steps emitted by the current Beacon 304 and the step of aligning the robot itself in a particular direction, using the next Beacon in the sequence, when including a transverse or remote room traversal sequence. Includes steps to move in that direction without using a guide from any of the above. FIG. 31G is a diagram showing the movement away from the navigation beacon 304 when the robot 302 finishes the migration.

During the migration, the robot 302 encounters its respective navigation beacon 304, while away from the base station 340, the coding of the first and second oriented beams 306 and 314, respectively. Store in memory. To return to base station 340, robot 302 performs coverage actions while turning on homing requests, resulting in finding the appropriate beams 306 and 314, respectively, and performing the corresponding coding to base station 340. Follow a series of encountered beams. In one embodiment, the robot 302 wirelessly (eg, radio frequency) communicates with the navigation beacon 304 to independently activate and search for each of the previously encountered beams 306, 314, and 316. It can follow each beam back to base station 340 (crossing the room using random bounces and / or obstacle tracking to place itself in the beam path).

In another embodiment (not shown), the navigation beacon is capable of emitting at least one, two, or three beam signals, which are fence / gate beams (as used herein). Called "yellow" but actually has the first modulation of the first frequency in the IR band) and the right navigation beam (called "red" but actually "yellow" In the IR band from the beam, with either or both of different modulations or different frequencies), and the left navigation beam (referred to as "green", but in practice "yellow" and "red" Has either or both different modulations and / or different frequencies in the IR band from the beam).

When multiple navigation beas are provided, each beacon has a different modulation of its beam, but the fence / gate beam, right beam, or left beam emitter is consistent within the navigation bea. Is preferable.

Regardless of its assignment as a fence, gate, or trail marker, as an example, navigation beacons are geometrically determined force fields in a circular area around their base (force fields are visible or IR). It is possible to have radio frequency RF instead of light, but it is also possible to radiate "blue" (which is actually one or both of the different modulations or frequencies in the IR band). It is possible to identify its presence to the robot, which allows the robot to follow the path around the navigation beacon, for example (via a gate or through a trail marker).

IR-based beacon systems can be designed to use geometric solutions, for example shaded floors that illuminate only a small area around themselves to provide a clear force field by IR. Similar to a particular perspective on non-limiting examples, such as lamps.

A fence navigation beacon, for example, emits a yellow beam directly in front of the navigation beacon (0 °) and a single navigation beam (red or green) at a 60 ° angle to the right or left of the yellow beam. It is possible. The gate navigation beacon is capable of emitting a yellow beam in front, a red beam at -60 °, and a green beam at + 60 °, as another non-limiting example. Also, for example, a trail marker navigation beacon can emit a red beam at about -60 ° and a green beam at + 60 °.

In some non-limiting examples, the "beam" of the navigation bea does not need to focus on a narrow area (and therefore does not require a special lens) and the beams do not interfere with each other. Practically, it is possible to fan the light over most of the room. Alternatively, for example, it is possible to use standard home-based LEDs that do not require other optics.

With reference to Figures 24A, 24B, and 32A to E, the docking behavior priorities (from highest to lowest) are docking retry docking bump tracking 264c, docking bound 264d, docking sharp turn 296c, and docking remote homing 296b. And the docking robe follow-up 296a and the docking field follow-up 296d. Robot 302 generally must approach dock 340 from the front with a skew angle of less than 4 ° in order to dock properly.

The default docking action, docking lobe tracking 296a, is performed when no other docking action has a higher priority, as shown in Figure 32A. When the robot 302 detects only the left (red) beam 350, it follows the edge of the red beam 350 in the clockwise direction. When the robot 302 detects the right (green) beam 360 or the overlapping region of the red beam 350 and the green beam 360, it follows the edge of the green beam 360 in the counterclockwise direction. This would cause the robot 302 to follow the outside of the nearest docking beam 350 or 360 around the front of the dock 340 at 0 °, and then follow the normal (0 °) edge 362 to the right of the docking contact. When the robot 302 detects the force field 370 during this action, it slows down to follow it more accurately.

Robot 302 records the signal detected by the omnidirectional receiver 322 through a 30-second window. When the robot 302 detects that the frequency and variance of the pre-crossing event (which intersects the edge 362 of normal (0 °)) is greater than or equal to about 1 Hz and the variance is less than or equal to about 2.5 seconds, the robot 302 normally Simple by following edge 362 (0 °) and performing smooth homing action under lobe-following action 296a (with predictable left-right amplitude) instead of continuing to follow edge 362. Judge to drive in a straight line.

Referring to FIG. 32B, the robot 302 encounters the docking beam 350 or 360, while being driven by a straight line, spiral, bump follower, or wall follower, not present in the force field 370, and the last number. The steep turn 296c is activated when the dock 340 is not detected by the directional receiver 318 per second. When the robot 302 detects the red beam 350, it moves clockwise toward the dock 340 in an arc. When the robot 302 detects the red beam 350 or the overlapping area of the green beam 360 and the green beam 360, it moves in an arc counterclockwise toward the dock 340. Robot 302 aborts this action 296c when it detects another beam 350 or 360, detects dock 340 with a directional receiver 318, or moves in an arc over 360 °. Generally, it is then followed by the lobe-following action 296a.

Referring to FIG. 32C, remote homing behavior when robot 302 has not detected force field 370 in omnidirectional receiver 322 in the last few seconds and has detected dock 340 by directional receiver 318. 296b is activated. When the robot 340 detects the red beam 350 or the green beam 360, it is driven toward them. If only the force field 370 is detected, the robot 340 is driven toward the force field 370. This allows the robot 302 to approach the dock 340 from a distance at any angle. When the robot 302 detects the force field 370 by the omnidirectional receiver 322, it stops this action 296b. During a frontal approach, this is generally followed by robe-following action 296a. During a lateral approach, this is generally followed by field-following action 296d.

Referring to FIG. 32D, the robot 302 has detected the force field 370 by the omnidirectional receiver 322 and has not detected the red beam 350 or the green beam 360 by the omnidirectional receiver 322 in the last few seconds. Also, the field-following action 296b is activated when the directional receiver 318 does not recognize the red beam 350 or the green beam 360. Robot 302 detects a red beam 350 or a green beam 360, bumps, cliffs, or moves more than 1 m in a random direction (clockwise or counterclockwise) by an omnidirectional receiver 322 to force field 370. Follows edge 372 of. When the robot 302 detects the red beam 350 or the green beam 360, bumps are detected because it moves straight away from the dock 340 at a distance of about 1 m or along the normal (0 °) edge 362. Until then, perform reverse lobe tracking 296a. The robot 302 then turns 180 ° or until it faces the dock 340. Operation 296b is then aborted. In general, the lobe-following action 296a is activated after the end of docking.

Referring to FIG. 32E, the robot 302 provides the robot's physical approach to the dock 340 from some approach angles due to the execution of the docking beams 350 and 360, or all, but not all, of the obstacle docking action 296e. Can be docked well, even when completely or partially blocked. Robot 302 is searching for dock 340 and also activates docking bound action 264c when it detects a docking beam 350 or 360 in the omnidirectional receiver 322 in the last few seconds and then also detects a bump. .. With a probability of about 66%, the obstacle docking action 296e starts bump-following action 264c, and with a probability of about 33%, it starts docking bounding action 264d, which simply backs up the robot 302, against lateral impact. Turn at a random angle of 10-170 ° or at 75-255 ° to a frontal impact, then stop action 264d. The bump-following action 264c follows the edge of an obstacle by the robot's bumper until one of the following conditions occurs: a) Robot 302 has an approximately 66% chance of detecting the dock 340 with the directional receiver 318 and not with the omnidirectional receiver 322, b) Robot 302 has an approximately 66% chance of detecting the dock. Detects the intersection of the normal (0 °) edge 362 from right to left and performs bump tracking 264c clockwise, or the robot 302 has a normal (0 °) edge 362 from left to right. Detecting intersection and performing bump tracking 264c counterclockwise, c) Robot 362 detects force field 370 by omnidirectional receiver 322, d) 30 seconds from the start of bump tracking action 264c Elapsed or e) Docking beam 350 or 360 was not detected by the omnidirectional receiver 322 for more than about 5 seconds.

Another possible implementation of the bump includes a step in which the robot 302 records the projection position of the dock 340 and a step of selectively turning towards the estimated dock position during bump follow-up 264c. Each time the robot 302 intersects the normal (0 °) edge 362 and detects the dock 340 by the directional receiver 318, the robot 302 uses odometry to be about 5 feet (around 5 feet) along the direction of travel. Approximately 150 cm) Project the position of the dock 340 over there. Robot 302 uses odometry throughout the docking maneuver to estimate its own orientation with respect to the projection position of the dock 340.

By a temporary combination of bump tracking and conventional docking methods, Robot 302 can dock in front of a wide variety of obstacles, including but not limited to walls, chairs, boxes, and reflections of IR docking beams. Can be done.

U.S. Pat. No. 6,594,844, entitled "ROBOT OBSTACLE DETECTION SYSTEM," discloses proximity sensors such as cliff sensors and wall-following sensors. US Pat. No. 6,883,201, named "AUTONOMOUS FLOOR-CLEANING ROBOT," discloses in detail the overall structure of the iRobot Roomba coverage / cleaning robot, as well as the main and edge cleaning heads. U.S. Pat. No. 6,809,490, entitled "METHOD AND SYSTEM FOR MULTI-MODE COVERAGE FOR AN AUTONOMOUS ROBOT," discloses behavioral control and coverage behaviors, including escape behaviors, selected by arbiters based on behavior-based robotics principles. doing. U.S. Pat. No. 6,781,338, named "METHOD AND SYSTEM FOR ROBOT LOCALIZATION AND "CONFINEMENT" discloses a robot confinement using a virtual wall, that is, a wall simulation directed beam. Each is incorporated herein by reference in its entirety.

Details and features of other robots that can be combined with those described herein can be found in the US patent applications filed with them. Application number _______, name "COVERAGE ROBOT MOBILITY", application number _______, name "MODULAR ROBOT", and application number _______, name "ROBOT SYSTEM". The contents of the above-mentioned application are incorporated herein by reference in their entirety.

A plurality of examples have been described. Nevertheless, it will be understood that various improvements can be made without departing from the spirit and scope of the appended claims. Therefore, other embodiments are within the scope of the appended claims.