Wireless sensor network

0001The present invention relates to a network of distributed wireless remote sensor units that use a battery as a power source and report data from various physical locations.

0002Such remote sensor networks are well known and are used to report a wide variety of applications such as temperature, pressure and the like. Such remote sensor networks provide a way to collect data from many different locations and centrally record, analyze and use it in control systems.

<p num="0003"> In many applications, the purpose of remote sensor networks is to collect data for medium- or long-term monitoring and performance analysis, so it is acceptable to have a relatively low update rate. It is possible. However, timing may be important from a safety standpoint. For example, in a gas sensor network for detecting abnormally high levels of dangerous gas, it may be crucial to report as soon as possible when the level rises to dangerous levels. .. This can be achieved by increasing the frequency of updates, but at the cost of significantly increased power consumption, which shortens battery life and increases cost and weight. It will be.</p>

<p num="0004"> The present invention attempts to address this issue. According to the first aspect, the wireless sensor network comprises at least one sensor unit having a sensor for detecting parameters and at least one query node, the query node issuing a query message. It is configured to periodically send to and receive a response message from the sensor unit, which has a predetermined deliberate delay unless the sensor unit determines that the alarm conditions are met. It is configured to send a reply message after, and if the sensor unit determines that the alarm condition is met, it is configured to send a reply message before the end of the intentional delay.</p><p num="0005"> Therefore, as will be apparent to those skilled in the art, according to the present invention, the remote sensor will be updated relatively infrequently under normal conditions (due to an intentional delay in responding to inquiry messages), but will have alarm conditions. When satisfied, it is possible to report more frequently. This has the advantage of extending the life of the battery for the remote sensor unit while preserving the capabilities used in safety-critical systems. The advantage is that the response is intentionally delayed under normal conditions, but the response is sent earlier under alarm conditions. This is a protocol that allows the sensor unit to control when it sends a response, but the remote sensor unit can only respond to the inquiry message, that is, it cannot initiate communication. Because it is suitable for.</p><p num="0006"> As used herein, the term "intentional delay" means a delay that is longer than an unavoidable delay in a system due to processing delays, potential clock latency, and so on. Is.</p><p num="0007"> Needless to say, the present invention also relates to a sensor unit of such a network. Therefore, according to another aspect of the present invention, the sensor unit for a wireless sensor network comprises a sensor for detecting parameters, periodically receives an inquiry message from an inquiry node, and an alarm condition is satisfied. Unless the sensor unit determines that the response is sent to the query node after a predetermined intentional delay, and if the sensor unit determines that the alarm condition is met, before the end of the intentional delay. It is configured to send a reply to.</p><p num="0008"> Further, according to the present invention, the method of operating a wireless sensor network having at least one sensor unit and at least one inquiry node is that the inquiry node periodically sends an inquiry message to the sensor unit and an alarm. Unless the sensor unit determines that the condition is met, the sensor unit sends a response to the query node after a predetermined deliberate delay, and if the sensor unit determines that the alarm condition is met, the sensor This involves the unit sending a reply before the end of the intentional delay. The inquiry node itself may be designed to create an inquiry message. However, in one set of preferred embodiments, the query node is configured to carry query messages produced by other devices, such as controllers connected to a common network with the query device.</p><p num="0009"> Alarm conditions can take multiple forms. In one set of preferred embodiments, the alarm condition is the threshold of the parameter to be detected. In the gas sensor example, the alarm condition may be a threshold concentration. Therefore, this alarm condition is satisfied when a concentration slightly exceeding this threshold value is detected. However, there are several other options, such as, for example, the lower threshold of the parameter (eg, oxygen concentration) or the acceptable range of (eg, temperature). Further, the alarm condition may be based on a compound condition, that is, a number of parameters exceeding 1. In addition to or instead, the alarm condition may include a time element. For example, the alarm condition may be the rate of change of the parameter, the average value, being out of the threshold for a predetermined time, or any combination of these.</p><p num="0010"> The sensor unit may have more than one alarm condition based on the same or different parameters, for example if the sensor unit has a plurality of detectors.</p><p num="0011"> The reply message may take several different forms. It may be the same under the normal state and under the alarm state. The inquiry node determines whether or not an alarm state exists by focusing only on the response timing. However, this is not appropriate for safety-critical applications. In a preferred set of embodiments, when an alarm condition is detected, the response message consists of a different message called the alarm message. This is to ensure accurate and prompt interpretation of alarm messages by the network and appropriate actions such as alarm generation. This alarm message may have a single flag or, for example, a multi-valued status code if the sensor unit has multiple alarm conditions. You may.</p><p num="0012"> The reply message may have quantitative data on the parameters detected by the sensor unit. Such quantitative data may form part of the response message under normal conditions, under all conditions, or only under alarm conditions. The data provided under the alarm condition is not absolutely essential, but may relate to the parameters that define the alarm condition. For example, if the alarm condition is an overconcentration of gas, the alarm message may include data about the actual concentration detected.</p><p num="0013"> According to the present invention, when an alarm condition is detected, a single alarm message is transmitted. However, preferably, the message is repeated and sent more times than a normal reply message. This ensures that the alarm message will eventually be received even if one or more packets are lost during transmission. It is also advantageous that the alarm message has quantitative data. This allows the data to be updated frequently, allowing the user or monitoring software to track the progress of the problem. Depending on the protocol, messages sent by the sensor unit other than reply messages, that is, messages that do not respond directly to request messages, may be treated as unsolicited messages and are therefore received by the query node / controller. , Is unlikely to be processed. Nevertheless, sending unsolicited messages is considered beneficial.</p><p num="0014"> In one set of preferred embodiments, the reply message has identifying information to identify the sensor unit. However, this is not absolutely essential. In this application, it is only necessary for the network to be informed at which point in the network the alarm condition is occurring so that corrective action can be taken, such as shutting down the system. is there. Alternatively, in systems where safety-critical is not particularly important, the query node may be able to infer which sensor unit is sending the response. .. This can be done, for example, by associating the sensor unit with a particular channel (time, frequency or code split), or by configuring the network so that the sensor unit responds only to inquiry messages sent to it. This makes it possible. In a preferred embodiment, individual query nodes can communicate with multiple sensor units, but this is not absolutely essential, even if individual query nodes can communicate with only one sensor unit. Good.</p><p num="0015"> Preferably, one or more sensor units are configured to relay messages and / or data from other sensor units or network nodes. This allows for a hierarchical network in which one sensor unit can act as an intermediate node for another sensor unit. This means that response messages from some sensor units go through multi-hop paths.</p><p num="0016"> The conclusion drawn from this is that the query node may have a sensor to detect the parameters. By doing so, the query node itself can detect the alarm condition and report it without needing to communicate with further query nodes.</p><p num="0017"> Preferably, the sensor unit is battery powered. This makes it possible to increase the flexibility regarding the place where the sensor unit is arranged. The sensor unit need not be powered by only one or more batteries. The sensor unit may further have a solar cell or other environmental power source to extend battery life. Multiple sensor units may share a single power source.</p><p num="0018"> In one set of embodiments, the sensor network according to the invention has multiple query nodes connected to each other and / or to a central server by a data network. Traditional wired or wireless data communication networks such as Ethernet, WiFi, TCP / IP, etc. may be used. In one set of embodiments, PROFINET and PROFISAFE are used.</p><p num="0019"> To facilitate this, the query node may be powered using mains power from the outlet. The central server may have a query node that is integrated with the central server.</p><p num="0020"> According to the present invention, the sensor unit is configured to send a response (eg, an alarm message) when an alarm condition is detected, faster than it would have thought if the alarm condition was not detected. Has been done. This still allows for intentional delays. Preferably, the sensor unit is configured to send a reply message as soon as possible after an alarm condition is detected.</p><p num="0021"> Preferably, the intentional delay of the sensor unit is configured such that the response to the inquiry message is sent just before the next inquiry message is received. In other words, the intentional delay should be as long as possible without overlapping the next message. This makes it possible to maximize the achievable advantage of the present invention that an alarm message can be sent as a reply as soon as possible when an alarm state is detected. This also has the additional advantage that the sensor unit can enter hibernation between receiving an inquiry message and sending a response to it, and only need to exit hibernation once in each cycle. doing. In hibernation, the sensor unit only needs to monitor alarm conditions and perform other essential functions, and the circuits associated with the transmitter and receiver can be turned off.</p><p num="0022"> In one set of embodiments in which the inquiry message is transmitted in period T, the intentional delay is greater than 80% of T. In other words, unless an alarm condition is detected, the reply message will be sent within 20% of the second half of the cycle after receiving the inquiry message.</p><p num="0023"> Preferably, the radio sensor unit is configured to communicate with the query node via a radio frequency signal. However, in some embodiments, other types of signals, such as ultrasound and infrared, may also be good candidates for unidirectional or bidirectional communication.</p><p num="0024"> In one set of embodiments, when an alarm message is received, the query node is controlled (internally or by an external controller) to send the request message more frequently. In the above case, it is preferable to have the sensor unit send a control message toward the sensor unit to the inquiry node to shorten the intentional delay by the amount that the request message is transmitted more frequently. This makes it possible to ensure that the sensor unit sends reply messages (ie, messages that should be guaranteed to be received by the protocol) more frequently. In addition to or instead, if an alarm message has been received, the sensor unit may send additional unsolicited messages. Since the unsolicited message is not a reply message, there is no guarantee that it will be received by the query node or processed by the query node or controller.</p><p num="0025"> The request message may be sent more frequently, for example, for a period of time or until the alarm condition no longer exists.</p><p num="0026"> The sensor network may be based on any of several different protocols, such as wireless HART (Highway Addressable Remote Transducer), but in one set of embodiments, the sensor network is determined by the International Society of Automation. It is based on the ISA100 standard.</p><p num="0027"> The parameters detected by the sensor unit may be any of a number of different options, depending on the application. In one set of embodiments, the sensor unit has a gas sensor.</p>

0028<figref num="1">It is the schematic which shows the sensor network which concerns on embodiment of this invention.</figref><figref num="2">It is a figure which shows the timing of communication at the time of normal operation.</figref><figref num="3">It is a figure which shows the timing of communication at the time of the alarm state operation.</figref>

0029Next, with reference to the accompanying drawings, preferred embodiments of the present invention will be described for illustration purposes only.

0030First, referring to Fig. 1, it can be seen that the network of hydrocarbon gas sensor units 2 and 4 is shown. Each of the sensor units 2 and 4 is equipped with a gas detector and a 2.4 GHz wireless transceiver, designed to operate with the ISA100.11 protocol or the HART wireless communication protocol, and is connected to the antenna 3. Of course, these specific details are intended for illustration purposes only, and other sensors and / or communication methods may be used instead.

0031Some sensor units of sensor unit 2 are directly connected to a gateway or proxy 8 by a direct two-way radio link 6. The gateway or proxy 8 acts as a query node, which will be described in detail later. The other sensor unit 4 is wirelessly linked to the other unit 2 to enable multi-hop coomunication between the remote sensor unit 4 and the gateway 8. In this case, the directly linked sensor unit 2 acts as an intermediate node. The sensor units may all be the same and may or may not be configurable as intermediate nodes as needed. Instead, there may be some that can only act as remote nodes. Other connection forms and hierarchies are possible. Therefore, there may be more than one hop in the path from a given sensor unit to the main network.

0032The gateway 8 is connected to the Ethernet (registered trademark) network 10, and the controller 12 that controls the operation of the system is also connected to the Ethernet (registered trademark) network 10. This operation will be described below with reference to FIGS. 2 and 3.

0033As can be seen from FIG. 2, under normal operation, the gateway 8 sends an inquiry message to a given sensor unit 2 with a downlink message safeReq (x) requesting confirmation of the safety status of the sensor unit. The controller 12 is instructed to do so every 20 seconds. The sensor unit 2 receives the safeReq (x) message and goes into hibernation for 18 seconds. In hibernation, the radio transceiver circuit and all other non-essential subsystems are powered down. After 18 seconds, the sensor unit resumes hibernation, creates and sends a response packet safeRes (x) based on the last received message safeReq (x). This response packet safeRes (x) is received by the gateway 8 and transmitted to the controller 12. Immediately after this, the controller 12 sends the following message safeReq (x + 1). This process is repeated. Therefore, the sensor unit is designed to receive once and transmit once every 20 seconds.

0034Figure 3 shows the operation under the alarm state. According to this figure, the controller 12 sends the first inquiry message safeReq (x) through the gateway 8 and the corresponding response safeRes (x) is received after a little over 18 seconds. The next inquiry message, safeReq (x + 1), is then sent, with the alarm condition that the gas detector detects the concentration of hydrocarbon gas above a predetermined threshold (other alarm conditions are possible). When (a) is satisfied, the sensor unit 2 will be released from hibernation instead of staying in hibernation for the next 18 seconds. Within the next 2 seconds, the sensor unit 2 sends a response in the form of an alarm message safeRes (x + 1) to the safeReq (x + 1) inquiry message. Needless to say, this reply is sent faster than it should be. The response also has different information, including a flag indicating that the response is an alarm message. The alarm message safeRes (x + 1) is sent to controller 12. The controller 12 can take appropriate measures such as warning of danger and closing of the safety valve. After sending the responsive SafeRes (x + 1) message, the sensor unit 2 continues to send the unsolicited message Unsafe (). The unsolicited message Unsafe () has data representing the concentration actually measured by the detector. The Unsafe () message is repeated every 2 seconds to provide up-to-date information about the gas concentration until sensor unit 2 is reset. When the following inquiry message safeReq (x + 2) (not shown) is sent, the sensor unit is "safe", i.e. responsive, as before, assuming the alarm state is still persistent. After responding with safeRes (x + 2), which is the reply message of, continue to send a series of Unsafe () messages. This means that the controller 12

0035Therefore, in normal use, there is only one receive and one send in 20 seconds (more precisely, because it sends the previous reply just before receiving the next message). It can be seen that the sensor unit can spend most of its time in hibernation mode and consumes very little power. However, in the alarm state, the sensor unit 2 can send an alarm message within 2 seconds. This acts as a response to the polling / inquiry message so that gateway 8 can immediately receive and process it. Therefore, wireless communication is effectively set up even when bandwidth is allocated asymmetrically. Wireless communication is set up once every 20 seconds for downlinks and once every 2 seconds for uplinks. This configuration allows the detector to be certified by the SIL2 standard and maintain a very long battery life. As will be apparent to those skilled in the art, communication between the sensor unit and the query node is further configured to include timeouts, sensor identification and data integrity checks to ensure compliance with SIL2 safety standards. Has been done.

0036Since the polling cycle is set to 20 seconds, the sensor unit will not stop functioning even if two packets are lost in succession within the process safety time of 60 seconds.