System for a fixed and/or moveable system in particular in vehicles, for example in cars

(54) Name: System and device for temporarily related events in a vehicle (56) Published publications: WO Al 2004 015 945, EP A2 1 191 747 (47) Abstract:

A party and / or mobile system, e.g. belonging to or connectable to a car (truck), performs party functions of two or more units included in the system (E1, E2). The occurrence of the functions (F-F5) is arranged for sensing and fixing by one or more time-setting means, e.g. clocks, and / or event indicating means which are respectively included in one or more of said units which constitute module units in the system. This may then consist of or comprise one or more serial, distributed and protocol utilization systems. Each party feeling is arranged to form the basis of or cause functional problems in the system. The invention makes it possible to significantly simplify the hardware included in each system, use smaller bandwidths or utilize these more for efficient information transfer. In addition, the system hardware can be simplified.

<img file="SE528607C2_D0001.tif" />

528 607

SUMMARY

A fixed and / or movable system, e.g. belonging to or connected to a car (truck), performs the determination of functions performed by two or more units included in the system (E1, E2). The occurrences of the functions (F-F5) are arranged sensing and determinable by one or more time perception means, e.g. clocks, and / or event indicating means which are respectively included in one or more of said units which constitute module units in the system. This may then consist of or include one or more serial, distributed and protocol utilization systems. Respective determination is arranged to form the basis of or cause functional action in the system. The invention makes it possible to significantly simplify the hardware included in each system, use smaller bandwidths or utilize these more for efficient information transfer. In addition, the system hardware can be simplified.

528 607

The present invention relates to fixed and / or movable systems primarily in or to vehicles, e.g. car. The system works with a common or relational time base for indicating time for detectors or generated events in the system and / or to this connected or connected system.

The present invention also relates to a device for operating in fixed and / or mobile systems, primarily in or to vehicles, e.g. car, effect the determination of functions performed by two or more units (nodes) included in the system. The device may be attributable to arrangements for sensing, controlling, analyzing and / or simulating input devices.

It is previously known that in e.g. vehicles utilize systems of this kind and it can be referred to, inter alia, filed with the same applicant as the present patent application and obtained patents. In the respective systems, the message and information (data) transmissions take place by or according to protocols of a kind known per se, which may be of a standard type, e.g. USB, CAN, LIN, Ethernet, IEEE 802. llx, IR, WUSB, etc.

In the related types of system, there are problems in determining events and / or time functions without relatively complicated and bandwidth-closing arrangements. It exists, for example. wishes to be able to determine the occurrence of the events and / or specify timing indications so that normal traffic can be utilized per se to indicate current time and / or event occurrences eg. without the need for additional hardware. The object of the invention is to solve, inter alia, this problem. There is a need to be able to work with preferably small resources in each module unit in order to simplify the structure of the units. There must also be great freedom of choice when designing systems and their interrelationships. The invention also solves this problem.

528 607

Any system in which the invention can be applied can be seen as individual, one or more series or sequences of events that are related to each other in time and space. Different time descriptions and time perceptions can exist within the same system. Smaller systems can be included in the system which, in turn, can now or later be included in other systems. During the first stage, the desktop stage, of a system development, it is advisable to relate all events to one and the same perception of time, for example related to the physically defined second. How well coordinated the various events must be in order for the desired system function to be achieved can be analyzed in a first stage. In a second stage, an analysis of the relation of the individual events to each other can at hand indicate that other temporal perceptions and utilization of the knowledge of the links of different events can simplify the construction of the final system and its description, verification and validation thereof. The invention simplifies these development processes.

When analyzing or verifying systems, there is not only a need to timestamp events that have occurred, but also with what precision and / or accuracy the timestamping takes place. The invention solves this problem.

In systems it appears that different parts of the system relate events to different time bases which in turn are related to each other. The invention facilitates this relationship position.

Analysis and monitoring instruments of various types for vehicles are often based on standard computers with operating systems, such as a PC with Windows XP. The operating system simplifies the development of the software to price of precision of timestamping of external events. Therefore, special units are introduced between the PC and the bus system. includes a clock for timestamping of incoming time message. For example, using the CANalyzer program from the company Vector, CANcardXL is connected from the same company to the PC. CANcardXL has a local clock and can timestamp messages from two CAN buses. If there is a need for several CAN buses, then another CANcardXL unit must be connected to the PC and the two CANcardXL units between themselves with a coordination unit via

528 607 coaxial cable arrangements to synchronize the local clocks to one another at a common time. With the practice of the invention, time stamping can be done with great precision, only used in the standard components of the PC.

What can essentially be considered a characteristic of a system and a device according to the invention is apparent from the characterizing parts of the subsequent main claim and the independent claims.

The occurrence of a functional practitioner may be arranged sensing and determinable by one or more time perception means, e.g. clocks, and / or event indicating means which are respectively included in one or more of said units which constitute or form module units in the system. The system may thereby consist of or comprise one or more serial, distributed and protocol utilization systems. Each party shall be provided with a basis for or cause malfunction in the system and / or in a system monitoring, analyzing, verifying, fully or partially simulating or stimulating device.

One consequence of the idea of ​​the invention is that one can easily time-relate the same or different events that are perceived or generated by different modules in the same or different systems. The modules that are thereby affected by the time relation may perceive certain events at the same time or alternatively certain events where the timing of these events is known to one another. Such events belong to the reference group of events. The modules may be arranged to generate or detect reference events, display local clock function, be able to read local time upon detection or generation of a reference event, and provide time information based on the read value transmitted to one or more translation units. The translation unit (s) then utilize these values ​​to create a common time reference for the modules and events in question, resulting in the creation of time entries that are mutually comparable. The translation unit can be considered to create a common time domain for relational events. Comparison can be done in different time domains that can be created by different translation units, which can use common reference events. In one embodiment, the bit pattern is used in one

528 607 communication protocol as reference transmission. For example, the Start Of Framebiten (SOF) in CAN is suitable. More specifically, the sampling point in said bit can be used because some CAN Controllers, for example Mcp 2515 from Microchip, generate a signal at this which can be used to trigger the required electronics to read local time for the event to occur.

Further developments of the inventive idea are apparent from the following subclaims.

So e.g. the system works with event functions and can thereby utilize reference events that can be related to one or more reference event generators. A reference event can be sent from a common point and be attributable to a common group. The reference events can occur at specified intervals that should in themselves be able to define clock function (compare Phase Locked Loop). The reference event detectors must be able to be placed in the units and thereby a detector can be placed in each unit. As a detector, interrupt generated by the selected protocol's communication circuit can be utilized. Each detector can be customized by defining or implementing the detector in the detector (eg connecting to the appropriate layer in the protocol stack and looking for special packet / bit pattern / flank). Events should be able to be grouped together, e.g. in terms of sequence or time. The group can work with a common starting point. So e.g. a device can communicate directly with everyone in the group or at least send messages that all other devices can detect. The messages can represent typical events. For example, each message may be packets according to some serial communication standard, compare the protocols mentioned above. The reference event may be wholly or partially a message. As a reference event, it is advantageous to select the USB protocol SOF package. This packet is normally transmitted and has sequence number, which facilitates proper association of timestamp to reference event. USB packages propagate in a USB system in a defined way and a USB host (host) can be seen as a reference event generator in its own time domain, different from other time domains, for example the domain an application under Windows works in, in the same PC where the USB host appears. The reference events can be broadcast to anyone who is part of the same USB arrangement, ie. all devices can

528 607 listening to the same event e.g. at the same time. By this is meant here a maximum of 50-100 ns jitter in the detection for all units plus up to a few hundred ns constant delays. A USB Hi-speed maximum delay can be 26 ns in the cable, 4 ns in "hub trace", 36 hs bits in the hub electronics in a maximum of 5 levels plus 30 ns for connecting the last unit, a total of 530 ns. A Hi-speed jitter due to the USB protocol can be a maximum of 5 hs bits per hub and a maximum of 5 hubs can be connected in a tree, which gives a maximum uncertainty of 25 hs bits, which corresponds to about 50 ns time inaccuracy for the propagation of a SOF through a USB tree.

In each unit, in said embodiments, a local clock may be included which may be readable by or in the application in question. Timing can be done by any reference event and the timestamping is preferably done with an externally triggered capture register. Correct association can be ensured by the timestamp with reference event. The timestamps can form the basis for the relationship setting of clocks, time and events. The clocks or the read time must be relational. In addition, an embodiment may include a time master whose time may be considered global (by group). In order to create a transitive relationship, that is, even if two different units cannot directly relate their time to each other, this can still be achieved if both can relate their times to a used master / intermediate unit. The time master may in turn be synchronized or related to a second reference, e.g. GPS. This gives the system access to a correct physical second that can be used in physical calculations, e.g. speed, power, accelerations, etc. The placement of the time master can be done at a common point in the current system, be separate or be included as another role in another unit. The role or function can be relocated. In one embodiment, the system does not need to use any master unit, but the translation is done in an all-to-all function. Furthermore, it should be noted that a time master need not have his own physical clock but can build a virtual clock function by utilizing time stamps of reference events made by and announced from modules, with their own time domains, connected to the system. Knowing the relations of the reference events and the respective time domain to each other, the time master can transform the respective time task into its own time domain and specify

528 607 respective time assignment referring to their own time domain to other units inside or outside the system.

An important part of an effective use of the invention is the actual development process of the device. In a first stage, the events to be generated and detected as well as their relationships to a common thought time base are identified in terms of times and allowed intervals around them for the device to exhibit the desired functionality. In a second stage, events that can be reference points between different entities are identified to detect, initiate or generate other events within the desired time intervals, relative or absolute. In a third stage, functions are implemented in an actual design to achieve the desired functionality. It should be noted that reference events are defined at the system level and that these need not be known at the module level. A basic idea of ​​the invention is the view of time. Any interaction between given event patterns can be considered to have its own time domain, ie. the time perception is based on its own time slot generator (which can be linear or non-linear) and the time indication given as the number of ticks, all plus any parts thereof, according to the linear or circular model. The different time domains can be transformed from one to another. Circular model domains will act as cyclic processes and those with a linear limited model such as discrete intervals in a linear unlimited model. For general analysis tools, it is advisable to use a linear unlimited model with the physical second as the timing. Another basic idea of ​​the invention is the view of a system. The system according to the invention is regarded as a number of event functions which are coordinated to an overall unit, the sum of all event functions. These functions can be of two kinds, event generating or event detecting. The coordination takes place in time and space. Part of the coordination takes place mechanically, some is done through information exchange between electronic devices, some of the latter via serial communication, for example of the type CAN or USB, directly or in combination. The system S is described by the number of n subfunctions F as:

528 607 s = Y^F.

where Fj is the sum of the number of m events h, i.e.

m

<img file="SE528607C2_D0002.tif" />

In the first stage, the system is described in one or a few time domains suitable for describing and calculating its properties, for example the time domain T. Symbolically this can be indicated by:

S (7j = £ / 7 (r) m)

^F&lt;(T) = X^hrsA^T&gt;)

In the second stage, a number of events are identified that are jointly linked between functions of interest and which are suitable for spanning a time domain of the functions. These events are referred to as reference events. Non-linear time slots in time domains for event functions that are mechanically coupled are often appropriate. Furthermore, reference events that are linked on mathematically descriptive chains to different time domains are identified. On the basis of a generated event in a time domain, whose detection generates an additional event (in the same or another time domain), which in turn is detected in another time domain, during the development work, a mathematical transfer function between two time domains can be created, like a conformal mapping of one coordinate system to another. Such transfer functions are implemented as needed in the various units of the system. Symbolically, the system can be described:

<img file="SE528607C2_D0003.tif" />

<img file="SE528607C2_D0004.tif" />

where V is a transfer / imaging operator between time domains 1 and m

528 607

In a further embodiment, one or more time coordinators may be utilized. These can delegate the eventual role of time master and receive and send timestamps of reference events. The transmission can then be made to any translation units so that they can calculate and translate the time in question. In a further embodiment, some units perform the translation function in question themselves. The time coordinates can receive and send other timestamps to any translation units. In one embodiment of the invention, the time coordinator is arranged at a common point in the system. The invention also takes into account that time translators can be included. Time translators handle the translation of time for the units that do not perform translation themselves. A time translator can be placed at any location in the system, but placement in a common point is advantageous from an efficiency standpoint. Time translators can perform translation from one unit to another via a selected reference time or according to the all-to-all principle, ie directly from each unit to any other unit. A logical translation matrix can then be used. However, a matrix that can keep track of how translation should be done quickly from everyone-to-everyone can assume large dimensions and be resource-intensive to keep up to date (the complexity scales squarely), but as a result provides faster translations. The time translation is also possible to perform only by keeping the reference timestamps updated, but this results in more work per translation. The translation unit can keep statistics on how well related clocks are arranged and send this information as a parameter along with the measured value. The inaccuracy can be calculated and transmitted to each measurement value. The translation function can be performed with greater accuracy afterwards, ie. when another reference message has been exchanged. In connection with this, for example, the derivative for the current period can be used instead of assuming that it applies to the same derivative as the previous period. More generally, interpolation is used instead of extrapolation to indicate a value. This possibility is particularly important when analyzing and verifying a system's function. Briefly and concisely, it can be mentioned that each unit / module that has prepared access to the occurrence of reference events in different time domains can be arranged to perform translation of the time domains between whether the units / modules themselves are included in / represent one of the aforementioned

528 607 domains or not. In order for no ambivalence to arise as to which time domain a particular indication belongs to, protocols should be drawn up that indicate, for example, who translates what and to what extent a unit's incoming / outgoing entries should be considered to belong to the sending and / or receiving unit's time domain.

The aforementioned developments may include that the system units can physically synchronize their local clocks after some master time. This in turn increases and complicates the hardware of the devices, which in such a case may still need to be supported by software resources. In return, all clockwork can be done proactively, ie. when the clock is to be used it already shows time in the master's time domain and it is only reading it which can reduce the response time of the unit and thus justify the higher complexity. Each unit can figure out how to clock its clock using the master's timestamps of the reference events. Each unit can have the coordinator / translation unit calculate how jerking and offset compensation should be done by sending their timestamps of the reference events to the same and then waiting for answers.

The units can translate their time to another time before the time in question is broadcast. This does not require any additional hardware, but some software resources may instead be included depending on how the translation is to be available. Translation work takes place between production and consumption of the value. The unit can calculate how the translation should be done using the master's timestamps of the reference events. The unit can have the coordinator / translator calculate how the translation should be done by sending its timestamps of reference events to the same and then waiting for answers. The units need not be aware of time other than their own in terms of time relation. It is enough for the units to send their respective timestamps of reference events to the coordinator. This requires relatively small resources in the units. Translators are responsible for all translation / time relation. Translators are arranged with computational power in relation to the translation method.

Depending on what precision is needed when detecting a reference event within a system, a group of individual events can be viewed as one and the same

528 607 reference event. For example, SOF in CAN messages can be utilized as a reference event. A practical way to detect SOF in CAN is to utilize the sampling point in CANcontrol. Since the detection of SOF in each node then depends on the setting of the sampling point (which may be different in different nodes) and the distance of the nodes to the transmitting module, the timing of the detection will vary depending on which node is transmitting and the receiving module's setting. If the SOF of arbitrary message is used as a reference event, it is in practice a group of events that are used jointly. The precision of the reference event is obviously not sufficient to measure, for example, the delay between different messages from different nodes. This is possible if individual CAN messages are selected as reference events instead (if the SOF is to be a reference point, the message may not, except in certain special circumstances, have won the bus arbitration from a message with a lower priority). More precisely, it may be that specific flanks / bits occur in the different nodes that provide the opportunity to determine delays between units, and that these flanks / bits have been propagated in both directions relative to the delay to be determined. Consequently, specific flanks / bits which have been propagated as undisturbed as possible through the system for determining delays are accordingly selected. A flank / bit that most likely originates from only one transmitting unit may be one after the arbitration field in a CAN message (however, preferably not the ACK bit, which is transmitted by receiving units). At the system level, you know which CAN messages and nodes each send. By sending and receiving modules each time stamping one and the same message, and message being sent both ways through the system, the delays between each module can be measured and calculated into a common time domain.

A device, e.g. an optional one, in one embodiment may act as a master of time (with or without its knowledge). The units can work with the translation function for any time to be translated / related to the master time. Translation work takes place between production and consumption of the value. This may cause a delay in possible consumption of the same.

528 607

The translation function can be done by adding the offset between the first and second time to the time to be translated. With this method, little computational power is needed for each translation. Preferably, synchronization / relational positioning is often done with this method to minimize the impact of the respective clock's operation in relation to the other clocks. The translation can also be achieved through both offset and fixed frequency error compensation (operating compensation). Translation A_x to B_x (new and old refers to reference time stamps) can be done according to:

B_x = + (B ™ - B_0LD) * (A_x- A ^ VCA ^ - Α, µ)

Given a computer with limited resolution and / or computational accuracy and given that times A and B are already scaled to go about the same (the derivative between them about one), the following method can give better results:

Βχ = B_NCW- (A_x - A_New) + (((B ^ - A_New) - (B_Old - A_0LD)) * (A_x - - A

Both methods are analytically equivalent and are based on linear regression. When calculating is done with a computer, discrete values ​​must be used which entail limited resources for representation, which means that the order is important. Regardless of the method, consideration must be given so that the calculation does not overflow / truncate in an undesirable / unpredictable way. The latter method may be one way to facilitate this.

Thus, a great deal of computational power can exist per translation, but in return relationship relations do not need to happen as often to achieve the same precision. The fixed frequency error can be determined offline or the frequency error is measured online. Wiring delays can also be of a non-negligible order of magnitude which must then be included in the calculation. In one embodiment, negligible line delays are selected. In the offline case, the offset and / or frequency deviation can be measured once for all and the required constants are calculated and entered in the translation function.

528 607

In an embodiment where delays can be measured online, units may be arranged so that reference function exercises propagate through the communication medium in both directions relative to the units whose delay is to be measured, during which both units determine the occurrence of reference function exercises for some time. These determined occurrences can be used to determine said delay.

The translator can save information in each unit about how the clock behaved during the last session in order to be able to phase in the unit more quickly in the next functional phase and / or make the synchronization / time relation easier for the translator.

The more of the aforementioned resources, methods and / or functions that can be dedicated, the more easily predicted and / or easily used and thus perhaps even safer the utilization of clocks, time, time synchronization and / or time relation in the system.

The specified system can advantageously be applied to a vehicle, for example a car, truck, tractor, scooter, boat, vessel, aircraft, etc .. For the direct control system of a car, it is advantageous to have time domains with varying times to coordinate movements such as are connected to the mechanical function of the driveline. Then every so-called. The ECU (Electronic Control Unit) comprises a CPU which is controlled by an oscillation circuit which is independent of the movement of the mechanics, has an ECU of at least two time domains with the associated transformation operator. One and the same ECU can interact with several other ECUs that form groups operating in a common time domain for each group. The entire system car (vehicle) can have a linear finite time domain, which starts, for example, with the ignition key position on and which ends with the ignition key position off. In this time domain, the system behaves in a manner appropriate to drive the vehicle. When the ignition key is in the "off" position, the system operates in a different time domain, for example, a circular time where information is collected from any on-going module that periodically listens for commands from a wireless signal to open or lock the car's doors and / or control alarm functions. The car can be included in other systems that deal with traffic control, traffic monitoring, law enforcement monitoring, etc.

528 607 which sets completely different requirements, but which sometimes interacts with the car's direct control system. Here other time domains may be better suited, for example a time domain where the time varies with the car's position. In a monitoring system, a need for updating the car's position may be lower when it is in rural areas under normal conditions than in a densely populated community or near an accident site. This frees up bandwidth requirements in both the car's control system and the monitoring system. The present invention has a great advantage in that the development of systems that will later be included in larger systems can take place without coordination. Once the systems are to be coordinated, one goes through the respective system's event generating and event detecting functions. Since each system has many such, it is highly probable that you find events that can be utilized as reference events for the coordination of the systems required to achieve the desired result. If suitable events cannot be detected in the coordination analysis, appropriate or appropriate event generating / detecting functions can be created and implemented in one or both systems.

A presently proposed embodiment of a device exhibiting the features significant to the invention will be described in the following with reference to the accompanying drawings, in which: Figure 1 shows a schematic diagram at the system level of how various input module units and systems are logically connected; 1 included in unit 103 in detail, Figure 3 shows a simple schematic view of a sub-unit included in Figure 1

105, Figure 4 is a schematic view of a first unit 104 included in Figure 1; Figure 5 shows two possible ways of measuring delays between units;

528 607 Figure 6 shows how the USB protocol's SOF package can be used for time relation in a system with a computer eg a PC and a number of devices connected to it, Figure 7 shows a traffic system where the cars use local independent time domains that are coordinated with a traffic monitoring system with centrally coordinated, flexible time domains, and figure 8 schematically shows a traffic monitoring system.

In Figure 1, three different systems are symbolized by S1, S2 and S3 connected to a system S. The systems can be system types according to the type mentioned in the preamble, e.g. USB, CAN, Bluetooth, etc. The serial bus connections in the systems are indicated by B1, B2 and B3. The systems operate with protocols P1, P2 and P3 attributable to said standards. The systems include one or more first units 104, 104 ', 104 ”. In accordance with the invention, functional events or exercises in the systems are to be relational with respect to some kind of time type, e.g. linear time, circular time, virtual time, etc ..

The systems work with a number of functions. Any first function is indicated by F1 and refers to one or more reference times or master times to which the first units can relate. The times in question may be e.g. real and / or virtual. The time any clock function 352, 410, 457 (not marked in Figure 1) represents can be an example of a real ditto, which can be done with or without the unit's knowledge. An example of a virtual time is a calculated mean time of appropriate units time.

The systems may also include one or more other functions F2 which represent translation functions between different times in the system. Depending on the characteristics of the times, the appropriate method is chosen to arrive at the translation functions. If

528 607 for example, two times both represent linear / modular times with tick of the same order of magnitude, linear regression of said type is advantageously used.

A third function is indicated by F3 and is arranged to generate / represent reference function exercises that can at least be detected by the units that can be directly relationally set. This function can be arranged either in one or more own units and / or in one or some existing units. It is suggested that appropriate existing functional exercises in the system are identified and arranged to constitute the function. Examples of such suitable existing functional exercises in a USB system are SOF packets, which are additionally provided with sequence numbers which facilitate proper identification of specific packets. A proposed existing functional exercise in a CAN system that would be suitable for the invention is the SOF bit alternatively suitable flank / bit within a message and then preferably a flank / bit which falls after the identification field but before the ACK bit. In addition, if a flank / bit before the data field is F3, the transmitting unit Tref could be sent as data in the same packet to save time and bandwidth. In order to make the system more stable, a flank / bit in a message should not be accepted as a reference function before the entire message is validated.

The first function F1 can be arranged to basically any of the units mentioned, with or without its knowledge. However, the second function F2 must, however, know which unit (s) possibly represents the first function F1, provided that F2 wants to use the F1 in question. The second function may be arranged in one or more of said units. In this case, the units can be arranged to make translations from or to their own times. Alternatively, one unit may allow another unit to perform the translation in question. The first function F1 can in turn be synchronized or relational to another time, e.g. UTC and the physical second via GPS, 107. Minutes can be drawn up for how times should be interpreted and / or translated. For example, if a group of units is arranged to be able to translate times between them, for example, incoming times that are already translated are selected because times that are first transmitted are translated to the receiving unit time or

528 607 vice versa. Alternatively, as mentioned Fl can be utilized in the form of both incoming and outgoing times being interpreted as / translated into said Fl.

In one embodiment, a second unit 102 can advantageously be arranged with the first function F1, the second function F2, the third function F3, the fourth function F4, the fifth function F5 and / or a sub-unit 105. Such an arrangement is particularly advantageous if protocol P2 is USB. 103 could in this case correspond to a unit 200 such as Figure 2 shows containing one or more units 201, where 201 may be internal and / or external USB hubs and / or devices such as 450. In one embodiment, the second function may be a centralized translator, which means that the first units could be performed relatively easily. On the other hand, decentralization of the translation function results in a slightly higher complexity of the first units, but in return the bandwidth may be chosen lower and fewer units utilized in the system, while reliability can be increased as resources are spread and thus no single point of failure. "Necessarily introduced.

Fig. 3 also indicates a fourth function F4 which detects the functions that the third function F3 generates. The first units 104 of FIG. 1 are provided with one or more sub-assemblies 105. FIG. 3 shows a simple schematic view of unit 104, where unit 300 corresponds to 104 and unit 350 corresponds to 105. 351 arranged to store times. At a current detection, the clock 352 is read and stored in a sub-unit 351 which saves the references in question, herein called Tref. The latter time stamping or reading can be transmitted to units performing the second function F2, which is thereby arranged to determine how the translation in question is to be performed. If the second function F2 is part of the unit itself, one or more steps from other or other units are expected, which respectively can be used to determine how translation to / from this and / or other units should be arranged to be performed. Alternatively, a unit's incoming Tref may be selected to synchronize the local clock to, for example, any F1, based on this or these timestamps. The Tref values ​​can be considered

528 607 is the basis for all time-related systems. This applies whether it is the issue of locally translating time, locally synchronizing the time in question, or transferring the translation to another unit. As can be seen in Figure 3, any of said functions / roles can be provided in unit 300.

An embodiment of the clock function 352 and the sub-assemblies 351 is to use a counter / timer in a suitable CPU and to propose a detection of F4 trigger a reading of said counter / timer 352 to a so-called capture register 351. The clocks are also intended to define system times and are readable by applications running in said devices, unlike the clocks hidden for applications, which are often used in certain time-controlled communication protocols. A very advantageous possibility offered by the invention is that clocks can be made as accurate as its applications require without, for that reason, forcing other clocks in the system to be equally accurate.

An advantageous accuracy can be obtained if the time of the first units is translated directly to each other instead of performing the translation via one or more first functions F1. More precisely, it can be said that negative effects arising from, for example, reading errors and jitter in the components and logic of input components can be limited. On the other hand, if you want to get all events in the system on a common time axis, for example for analysis, it is advantageous to use a first function F1 and mapping up current events along the time axis F1 described. A fifth function F5 indicated in Figures 1 and 3 consists of a possibly used coordinator function which receives and sends reference timestamps to the person who wants or is to use the relevant timestamps, preferably to all the units working with function F2. Conveniently, the fifth function F5, with or without guidance from one or more other functions F2, determines which unit (s) should operate / hold the role of F1. In addition, the other functions of the system or systems could calculate and send with / associate the respective translation accuracy / inaccuracy and / or stability / instability of the times / clocks / translations. This in order to avoid assuming that the worst case always prevails but instead be able to make use of prevailing conditions.

528 607

In the most general case, the units do not need to know who / what / which is the origin of a specific reference function. However, if the units are arranged to be able to measure any delay over the serial communication between two units, they should be given the opportunity to do so. Said delay can, for example, be determined with knowledge of a number of occurrences of reference function exercises propagated in both directions relative to the units whose delay between them is to be determined. Figure 5 shows examples of how this can be done.

For example, if said unit 102 is a computer equipment type PC with an operating system such as Windows XP, the functions F1, F2 and / or F5 could be implemented in software that communicates with the OS interface to the communication channel. The advantage of this is that PC computers are very common today and are often already included in many intended target systems and also usually have resources over which could be advantageously utilized. Depending on the implementation method and system requirements, great consideration must be taken so that calculations etc. are done with a sufficiently short response time, regardless of what the OS is otherwise doing. This embodiment can provide a hardware simple, flexible and thus cost-effective way of connecting, for example, a computer to multiple CAN systems for analysis and / or interaction and at the same time obtain a competent way of relating time in and / or between the systems. Figure 6 shows examples of how this can be done.

Figure 4 shows in greater detail than Figure 3 a schematic structure of a unit 104 according to Figure 1, designated here 400. For the sake of clarity, it is provided with two microprocessors, but the task can be solved with a microprocessor. The unit 400 is connected on one side with a system 401 via the contacts 402, 403 and the connection line 404. Via the matching electronics 405, the signals on the bus can be read by the microprocessor 406. By means of instructions stored in memory 407, the signals can be interpreted in accordance with the simplest form of protocol 408.1 in the system, the interpretation may mean that only the received bit pattern is transmitted, but the interpretation can be of a comprehensive nature where much of the additional information given by the rules of the protocol is supplied by microprocessor. Application software, ie instructions for one or more applications processing for the microprocessor

528 607 available information is also stored in memory 407. The information so interpreted is transferred to the double ported memory 409. Additional information of interest may be added to the interpreted information, for example, time stamping when the information was obtained from the system. The timing is retrieved from the clock 410 which is triggered to be read appropriately by the adaptive electronics 405, 414, 419 or function exercise detectors 424, 426, 428, for example, when reception of a message is initiated. In order to make the reading of the clock more independent of the processor 406, the occurrence can be stored in capture register 425,427 and / or 429. The information is stored in the double-ported memory in an organized manner according to a set of rules adapted to the requirements of the system protocol so that specific information is stored at a specific location indicated by table 411. The dual port memory 409 can be read by microprocessor 412 which can communicate according to a second protocol by means of rules stored in memory 413 and physically via adaptation electronics 414 with a second system 415 via contacts 416 and 417.1 memory 413 also stored rules for how the information is stored in 409 according to the rules of 407 and 411 are converted according to the rules of the second protocol 418.1 in simpler systems, the second protocol can be based on CAN and several units 400 ', 400, etc. of the type 400. In the same way as previously described, the unit 400 also contains rules for a third protocol with the adapter unit 419 and the connectors 420 and 421 which connect to the connection 422 with the protocol 423. A suitable protocol may be based on USB.

In order to be able to synchronize a local clock 410 to an external time, for example some F1, the processor should at least be able to set a new value in the clock. However, for the sake of simplicity and reliability, the interface between the clock and CPU should allow, for example, the clock to compensate for the clock itself for a given offset and / or frequency error in a manner acceptable to the unit and the system.

Since the possibilities of getting Windows to perform tasks in real time are severely limited, it can be advantageous to let 102 be a separate computer system with an OS better suited to the task. One may be a unit 450 of Figure 4 with a microprocessor with peripherals specially adapted to handle communication and computational problems. To the unit 450, a number of modules 400 can be connected via a suitable connection, for example a USB connection. Such an arrangement has

528 607 many benefits. The analog and hard real-time near problems are solved by module 400, while the computationally heavy and less real-time critical data is solved by unit 450. Module 450 is connected directly to one or more modules 400 with connectors shown with 451, 452, 45Γ, 452 ', etc. Communication circuits 453, 453 'are connected to a microprocessor 454 with peripheral equipment, e.g. memory (-n). These memories include application software, ie. instructions for one or more applications that process information available to the microprocessor. A memory card 455 is provided for logging equipment, recording and playback capability, etc. A memory 456 connected to the microprocessor is writable and readable from two (both) directions (i.e.). from both the system side and the tool side The memory can be divided into a number of sub-memories with different algorithms 456 '. The clock 457 can be synchronized / relational with / to the clock of the first unit 410 via the protocol in the above manner. As a help, there may be function exercise detector 463 and capture register 464 arranged in a similar manner to unit 400. Thus, all first units 104 may be connected to a second unit 102 duration-synchronized / time-relational. Similarly, other units could relate to one another. Through time synchronization of the different units, the execution of the applications in the different units can be synchronized or related to each other. Execution of applications or application parts responsible for measurement can thus be coordinated with the execution of applications or application parts responsible for communication within and between the different units. This means, among other things, that messages sent according to an event-controlled protocol, for example CAN, can behave in a time-controlled manner because applications for sending messages are executed in a coordinated manner over time. By coordinating the execution of applications for measurements with the transmission of measurement results, a time relationship is obtained between the measurement and the distribution of the measurement result in the system in the form of messages. The same can of course be done for indicated events and messages with information about each event. The unit 450, together with the unit 400, can also simulate fully or partially an ECU in an ordinary CAN system in a vehicle, even with suitable software. Unit 450 may be equipped with means for communicating with other network protocols, for example Bluetooth 459 and TCP / IP 460 for communication between a network of units 450 and / or tool unit implemented in PC or PDA. As an alternative to memory disk, so-called "USB mass storage devices" can be connected to a USB connection. For communication over telecommunications networks, the unit may be equipped with a GSM module 461 and for time synchronization

528 607 or clock synchronization with a GPS module 462 which can also be used for position determination. Regarding the said protocol, cf. above.

Regardless of whether unit 102 is an ordinary PC or a unit such as 450, it can advantageously perform the function F3 and especially if the selected protocol is USB and the unit is USB host, which means that the unit generates SOF packets. These, as previously mentioned, are suitable to constitute reference function exercises.

Figure 5 shows the possibility of measuring the communication delays between units, where units 501 and 502 can be of said type, for example as unit 400.1 in the first case, unit 501 sends a message 503 to unit 502 and both timestamps the message with its local clock. The 501 timestamp of 503 can be called Trefl3, and 502's timestamp of 503 Tref23. Thereafter, 502 sends a message 504 to 501 which is also timestamped by both units. 501's timestamp of 504 can be called Trefl4 and 502's timestamp of ditto Tref24. For example, if 502 then sends said Tref23 and Tref24 to unit 501, 501 may determine the communication delay, Τ_Λ = Γ_Λ50,+7^+7_Λ502, between 501 and 502, for example, by the following method:

<img file="SE528607C2_D0005.tif" />

A further variant for determining delay is also found in Figure 5. If instead, a unit 505 sends a message 507 which, in accordance with the figure, first reaches 501's switching point on the communication channel and then 502's, and 501 timestamps the message as Trefl7 and 502 as Tref27. . Subsequently, unit 506 sends a message 508 propagating through the communication channel in the opposite direction. This message first reaches the 502's switch-on location and then 501's. 502 timestamp it as Tref28 and 501 as Trefl8. The delay on the common portion of the communication channel 7 ^ can then be determined as follows:

<img file="SE528607C2_D0006.tif" />

528 607

Both of these simple variants of delay determination presuppose that Trefama uses time sticks in the same order of magnitude and that any frequency deviation between units 501 and 502 can be neglected or compensated for by any of the aforementioned methods.

Figure 6 shows an embodiment of how the USB protocol's SOF packet can be used for time relation according to the invention. The system includes a number of units 603, 603 ', 603 "which may be of the said type for example 104 or 400. These are connected via one or more USB hubs 602 (for example of type 103 or 200) to a computer 601 for example a PC as well as it can be of the said type, for example 102. 601 transmits itself according to the USB protocol SOF packet 651 at regular intervals. This can be an example of the aforementioned function F3, ie the reference function exercise generator / reference event generator.

A Start-Of-Frame package, SOF, typically contains a Start-Of-Packet / Sync field 691, an identification field 692, the identification field inverted 693, front number 694, CRC field 695, and a final End-Of-Packet field. field 696.

603 contains a USB controller 631 which in turn contains a device 632 intended to detect SOF packets 651. The detectors of 632 are arranged to trigger a reading of the clock 634 to a capture register 633. The time thus saved is hereinafter referred to as Tref. Both clock and capture registers are readable by microprocessor 635. Both 633 and 634 could advantageously be built into the 635.1 memory 636 is stored program code which is run, for example, every time 635 receives a SOF packet from 631. The program can, for example, read the frame number of the SOF packets and depending on the requirements and conditions always or at a certain interval read out Tref from 633 and send in a new USB package 652 using 631 via 602 to 601. 603 contains also adaptation electronics 637 to a CAN controller 638 which in turn contains a precise detection mechanism 638 'for when CAN messages start. The detection mechanism also triggers a capture register 639 which then reads 634. The timestamp contained herein in 639 can be sent together with the message that triggered the reading in a packet 653 to 601. The packet type 653 can e.g.

528 607 consist of USB overhead, 642 and 642 "", data, 642 ', from CAN controller 638, a timestamp, 642 ", of when data 642' was received by 638, and possibly more data with associated timestamps, 642 '" . The packet type 652 consists of, in addition to the overhead, 641, 641 "", which USB supplies by a Tref, 641 ', the sequence number of the SOF that gave rise to this Tref, 641 ", and possibly more Tref's with the corresponding sequence number and / or other data, 641 '”.

601 is provided with a USB host 671. As already mentioned, 671 periodically sends SOF packets 651 in accordance with the USB protocol. When the computer receives type 652 packets from its connected devices 603, 603 'etc. via the USB host 671, these can be read and processed by processor 672 by means of program code in memory.

The 673.1 processor runs, among other things, program code according to the logical data flow diagram in 680. 681 here represents the USB drivers that form part of the operating system that 601 uses. These, for example, keep track of the physical structure of the USB system and read out data packets from the controller 671. If these packets originate from any of the units 603, 603 'etc., they are forwarded via the time management functions 682 to the intended drivers 683. If the messages are packets of type 652, the information 652 'is picked from them, ie the so-called steps 641', ​​the sequence number 641 "and any other data 64Γ", and they are taken care of by the time management functions 682 which are part of the 683. The system is supposed to be used as well and it is represented here by a user application 685 which logically communicates directly with the units 603 etc. but practically it is done via the commonly known interface 684 which provides precisely all the functions that a user can use in the devices. The interface 684 thus interconnects the program 685 with the device drivers 683. For example, the interface presents, if desired, all timestamps from the different units from one and the same time axis, which is done by means of the time management functions 682 entirely in the spirit of the invention. In the figure, this is exemplified by the message 653 sent from a unit 603 to 601. Unit 671 decodes the packet and forwards the contents, 653 ', to the processor 672 for further processing. For example, unit 681 sees that the message originates from a unit of type 603 and forwards it to the intended drivers 683. The time management functions 682 of 683 see that the message contains a timestamp 642 "in a local unit's time and therefore translates the same

528 607 at a suitable time according to, for example, the following procedure. The translated time, 643 ', and possibly the calculated inaccuracy, 643 ", are presented together with data, 643, (ie 642' possibly further processed by 683) and any other info, 643 '", combined 653 ", for the application 685 via the interface 684th

The time management functions 682 can be divided into a part which saves and manages a history, 682 ', of Trefs with the same sequence number coming from the packets 652' as well as information about the origin unit and keeps the translation functions updated and a second part which performs the translations themselves. In order to perform direct translation between all units' times, a kind of logical translation matrix 686 can be used. For each pair of units 687, 687 'etc whose time can be directly translated there is a list 688 with most recently matched pairs of Trefiar 688', 688 'etc. These Trefiar 688' etc can then be used to perform a translation between the times of incoming units, for example. according to:

Βχ = B_mw + (B_New - B_Old) * (A_x- A ^) / ^ - A ,,,) alt ^Bx = B_NCW- (A_x - A ^) + (((B_New - A ^) - (B_Old - A_0LD)) * (A_x - A ^)) / ^ - A ^) where A_x should be translated into B_x using the matched Tref pairs (A_NCW, B_mw) and (Αθ_ω, B_Old) which can thus be equivalent to 688 'and 688' ". Apparently there are some calculations that can be done in advance, for example (B_New - B_Old) and a_{ne w} - A_Old) in the first option and (B_New - A ^ J and (B_Old - A_Old) etc in the second option. Such calculations can, as mentioned, be advantageously calculated in advance and saved in the designated location 689 in order to simplify and thus speed up future translations. The statistical inaccuracy in question can also be calculated and saved in the designated location 689 '. Of course, it is possible to make a very complex and exhaustive estimate of possible inaccuracy based on the entire history of matched Tref pairs, but to give a simple illustrative example, for example, one of the Tref's in a pair of 688 'can be translated with the help of two others. pair in list 688. The translated Trefi is then compared to the actual Trefi in the aforementioned pair. The difference between them can be seen as a simple measure of inaccuracy / non-linearity.

528 607

Another variant of the translation matrix may be to insert / allow a row / column to represent a virtual time based, for example, on some weighing of suitable other times.

A simpler version of the translation matrix is ​​the special case that only one or a few rows / columns in the matrix are kept updated in the above-mentioned manner. Since each row / column can give information on how relationship relations can be accessed from and to a certain time, this time can be used as a kind of master / intermediate time in translation between two other times. This may mean a less memory and / or resource-demanding alternative to keeping the entire matrix updated to the price of slightly more complicated translations and / or possibly greater inaccuracy. The master / intermediate time here may be examples of said function F1.

The translation function F2 / 682 can be broken down into a number of parts which, however, do not necessarily have to be performed in each unit arranged with the function. Some steps can be performed by a unit which then sends that information to another unit which then does not have to carry out the said steps. Examples of parts:

• is to collect and keep track of time-stamped reference function exercises for any length of time that can be translated.

• is to find timestamps from different units of the same reference function performing among collected timestamps.

• is based on these matched timestamps to determine how a possible translation should be made depending on given conditions.

• may be based on matched timestamps to determine how stable / accurate a particular time can be considered, for example by means of statistics.

• may be, if a time is considered unstable / inaccurate, to request more frequent time-stamped reference function exercises from the system.

• is to perform translations when requested.

The time management functions can be considered as an example of said function F2.

528 607

In the said system, for example, SI can work with local translation of time, S2 with centralized translation and S3 with local synchronization. Of course, it is also possible to mix different variants in one and the same system.

The event recordings should be arranged or form the basis of or cause functional measures in the system such as, for example, analysis. Such function measures are represented in Fig. 1 by F6 and F7.

Figure 7 shows a simplified car system 900 consisting of a number of event generating subsystems exemplified by a motor 901, a gearbox 902, four wheels 903, 904, 905 and 906 and the doors 907 and 908. Other ECUs for the car's functions required ECUs are symbolized by 900 '. Each event generating subsystem has an event generating and event detecting ECU (Electronic Controller Unit). The engine ECU 909, the gearbox ECU 910 and the wheel ECUs 911, 912, 913 and 914 are connected to the CAN network 915. The door ECUs 916 and 917 are connected to the LIN network 918. The LINmaster 919 is connected to both the LIN network 918 and the CAN network 916. and monitoring unit 920. External devices such as diagnostic instruments, programmers, analyzers, etc., represented by 921, can be connected to the system via connector 922 and a connection cable 923. Contact from external devices 924 can also be made via radio unit 925 and radio connection 926 in monitoring unit 920. In a first stage, the monitoring unit 920 commands all other modules in a first time domain where no time elapses before it sends a wake-up event over the CAN bus. In this mode, the system's energy consumption is very low. The monitoring unit itself is located in a second low-consumption time domain where a time event is generated by an internal clock 920 ', for example 10 times per second. This event generates eavesdropping of any incoming signals from an external unit 924, for example, a door opening signal. If such a signal is detected, the monitoring unit sends an awakening signal over the CAN house and all ECUs switch to a respective local time domain suitable for CAN and LIN communication and for internal communication.

528 607 event detection and event generation. Orders for door opening are given to unit 919 which in turn orders units 916 and 917 to open the doors.

The ECU's controlling rotating parts establish their own time domains based on the angular speed of rotation. In its simplest form, time base events are generated by pulse generating wheel 930 and pulse detecting means 931. In this example, a pulse with the reverse sign is generated by a sensor 932. The means 931 detects pulse train 933 which is passed to ECU 934 thereby creating a clock function 934 'with a circular time base with a time stroke corresponding to 22.5 degree angular rotation. Thus, time is counted locally in elapsed wheel angular units. If the wheel stops, the local time is still. The ECU 934 can control the brake 935. The time base varies with the wheel speed compared to a time base based on the physical second. A commanding unit 936 with a time base 937 based on the physical second can issue commands K1 and receive reports R1 from the unit 934. By position reports from the four wheels, the unit 936 can determine how the wheels move relative to each other and command them to increase or decrease the braking force during the lap. If a wheel locks, the time stands still and reporting stops. This condition is easily detected by the unit 936, partly because other wheels report, and partly because the expected report in relation to the internal time domain has not been received.

More complicated time domains can be built up when appropriate. The event generator 940 generates two pulses per revolution with alternate characters and 45 degree intervals. These are detected by the detector unit 941 and transmitted to the converter 942 in the ECU 943 via connection 944. The generated time domain is displayed as 945. The ECU 943 has an internal clock 946 which generates the time domain 947. The system includes a pressure transducer 948 which, via connection 949, provides pressure information in discrete steps to the unit 943, where each step can be seen as an event. Passage of some of these events can be utilized to generate time sticks. The pressure transducer is connected to a cylinder 950 in the engine 901. The ECU 943 controls the electric valve arrangement 951. The local translator 942 compiles the information from the time domains 945 and 947 and 948 into a new time domain suitable for valve control.

528 607

952nd The unit 953 provides control commands to the valve control mechanism 951 via the connection 954 in time with the time domain 952. The local translator 942 thus obtains a clock function seen from the unit 953. In this way, a virtual camshaft can be created where the control algorithm's properties are changed by changing the time domain 952 with speed. pressure variations.

For analysis or diagnostics, a unit 921 can be connected to the system. The different nodes may have stored in their minds a translation algorithm that specifies how their respective time domains are transformed into a linear time domain based on the physical second. The instrument 921 can request that each module transmit the translation algorithms. This assumes that it is used so-called. HLP (Higher Layer Protocol) supports such a procedure. Otherwise, the algorithms may be in the system documentation. Alternatively, the algorithms are stored in the system module 925.

Figure 8 shows a schematic traffic monitoring system. Traffic control and traffic monitoring are an increasing problem in society. More and more cars are equipped with GPS and GSM to make it easier for the driver of the car. This development is largely independent of the aforementioned social problem. A simple solution to the problem is to standardize a fixed local time domain for all cars and that the monitoring system works with time domains that vary geographically and with needs. The cars are required to be equipped with GPS receivers and transmitters to a communication network determined by society and report their position periodically, for example every ten seconds. If a crash occurs, this is reported immediately with data from common crash-related sensors. The road network is covered by communication cells 970, 970 ', 970', etc. of suitable size. Each cell is served by a base station 971, 97Γ, 971 ”, etc. connected to a wired (optical or copper) network 972. To this is also connected the control and surveillance system, symbolized by 973. A number of cars 974, 974A, 974B, 974C, 974D, 974E travel the road 975. Every ten seconds they transmit their position according to GPS exemplified by 976, 976 '. In addition to positioning, additional information can be sent, for example speed, in the event of damage estimated, number of passengers, etc., depending on legislation such as anonymous or identified sender. Vaije base station does a first machining and

528 607 compiling data, for example, the number of cars passed per minute, accident occurrence, violated speed limit, etc. and compiles this in a report 977. Each base station is assigned by the unit 973 at least two local time domains, for example a circular with run time 1 min and a linear limited , with start time for a special type of message received from a car, for example, crash message that extends even sent report to unit 973. In this way, the utilization of the road communication network is optimized. The traffic information comes regularly once per minute over connection 973, scattered over time because the local time domains are not synchronized between themselves. In the event of an accident, reporting occurs immediately and can thus be handled quickly. The transmitted message from the crashed car becomes a reference event between the car system and the monitoring system. 973 can change time domains as needed and the bandwidth of 972 can be made available for other information, such as commercial communications. The time domains of cars do not need to be related to the monitoring system. It will be readily appreciated that the notion of invention can be varied in many ways to control and change the characteristics and relationships of other systems to other systems of interaction between them.

The invention is not limited to the exemplary embodiment shown in the above but can be subject to modifications within the scope of the following claims and the inventive idea.

528 607