The present invention relates to a method and a device for adapting the data transmission reliability between at least two subscribers in a serial bus system.
The ISO standard family 11898-1 through -5, for example, describes a Controller Area Network (CAN) as well as an extension of the CAN called “time-triggered CAN” (TTCAN), referred to in the following also as standard CAN. The media access control method used in the CAN is based on a bit-wise arbitration. In bit-wise arbitration, multiple subscriber stations are simultaneously able to transmit data via the channel of the bus system, without thereby interfering with the data transmission. Furthermore, the subscriber stations are able to ascertain the logical state (0 or 1) of the channel while transmitting a bit over the channel. If a value of the transmitted bit does not correspond to the ascertained logical state of the channel, the subscriber station terminates the access to the channel. In CAN, the bit-wise arbitration is usually carried out on the basis of an identifier within a message that is to be transmitted via the channel. After a subscriber station has sent the identifier to the channel in its entirety, it knows that it has exclusive access to the channel. The end of the transmission of the identifier thus corresponds to a beginning of an enable interval, within which the subscriber station is able to use the channel exclusively. According to the CAN protocol specification, other subscriber stations may not access the channel, that is, send data to the channel, until the sending subscriber station has transmitted a checksum field (CRC field) of the message. Thus, an end point of the transmission of the CRC field corresponds to an end of the enable interval.
The bit-wise arbitration thus achieves a non-destructive transmission, via the channel, of those messages that won the arbitration process. The CAN protocols are particularly suited for transmitting short messages under real-time conditions, a suitable assignment of the identifiers being able to ensure that particularly important messages will almost always win the arbitration and be sent successfully.
With the increasing networking of modern vehicles and the introduction of additional systems for improving driving safety for example or driving comfort, the demands grow on the quantities of data to be transmitted and the latency periods admissible in the transmission. Examples are driving dynamics control systems such as, e.g., the electronic stability program ESP, driver assistance systems such as, e.g., the automatic distance control ACC, or driver information systems such as, e.g., the traffic sign detection (cf. for example descriptions in “Bosch Kraftfahrtechnisches Handbuch,” 27th edition, 2011, Vieweg+Teubner).
DE 103 11 395 A1 describes a system in which asynchronous, serial communication is able to take place alternatively via an asymmetrical physical protocol or via the symmetrical physical CAN protocol, and thereby a higher data transmission rate or data transmission reliability is achievable for the asynchronous communication.
DE 10 2007 051 657 A1 provides for the use of an asynchronous, fast, non CAN-compliant data transmission in the exclusive time windows of the TTCAN protocol in order to increase the transmitted data quantity.
G. Cena and A. Valenzano, in “Overclocking of controller area networks” (Electronics
Letters, vol. 35, No. 22 (1999), p. 1924) deal with the effects of overclocking the bus frequency in subsections of the messages on the effectively achieved data rate. The adaptation of data transmission reliability is not discussed.
It is clear that the related art does not provide results that are satisfactory in every respect.
The present invention is based on the transmission of messages being of a logical structure according to the CAN standard ISO 11898-1 in a bus system that includes at least two subscribed data processing units, the logical structure including a start-of-frame-bit, an arbitration field, a control field, a data field, a CRC field, an acknowledge field and an end-of-frame sequence, and the control field including a data length field, which contains an item of information regarding the length of the data field.
According to an example embodiment, for certain transmitted messages, the present invention provides an option to use a modified polynomial for calculating a checksum and to transmit a CRC field of a size that deviates from the CAN standard in that the CRC field of the messages may have any of at least two different numbers of bits depending on the value of an associated switchover condition. This makes it possible, as a function of the switchover condition, to adapt the data transmission to the respective data transmission task, for example to the scope or the security relevance of the transmitted data, and thus to influence or define the data transmission reliability. For this purpose, according to an example embodiment, advantageously at least two different generator polynomials are used for defining the content of the CRC field as a function of the value of the associated switchover condition.
Regarding the use of the method according to the present invention and the corresponding devices even in conventional CAN networks, it is advantageous if, for at least one value of the associated switchover condition, the number of bits in the CRC field and the generator polynomial used for defining the content of the CRC field correspond to the CAN standard ISO 11898-1.
In an example embodiment, the corresponding messages are detectable by a suitable identification in the arbitration field and/or in the control field. This allows the subscriber units receiving the message to recognize the messages modified in accordance with the present invention and to adapt their receiving process accordingly. This can be advantageous if the content of the data length code is also used for this adaptation.
In an example embodiment, multiple calculations of checksums are started in parallel at the beginning of a message and, as a function of the satisfaction of an associated switchover condition and/or the content of the data length code, a decision is made as to the result of which of these calculations is used or transmitted in the CRC field. This makes it possible to transmit the information as to whether a message is transmitted according to the standard-conforming method or according to the method modified in accordance with the present invention along with the message, without informing the recipient in advance about the used method. The checksums for checking the correct data transmission exist for both methods and may be evaluated as needed.
By providing a possibility of enlarging the data field of a transmitted message, an example embodiment of the present invention achieves the effect that, compared to a standard-conforming CAN message, a greater quantity of data may be transmitted over the bus in a single message. This advantageously increases the ratio of data quantity and control information in a message and thus also the average data transmission rate over the bus system. The combination with the adaptation of the CRC field according to the present invention has the advantage that the reliability of error detection is maintained or may be adapted even for larger transmitted data quantities.
By establishing an unequivocal correlation between the content of the data length code and the length of the data field, a high flexibility is advantageously achieved with respect to the possible size of the data field.
It is furthermore advantageous that, for the values 0b0001 through 0b1000 of the data length code normally used in standard CAN, the sizes of the data field corresponding to the CAN standard, that is, 1 byte through 8 bytes, are assigned and the remaining values of the data length code are used for the additional admissible sizes of the data field up to the maximum admissible size. This cost-effectively reduces the adaptation effort of application software when switching to the method according to the present invention.
The use of a modified polynomial for calculating the checksum occurs as a function of a switchover condition such that, when the switchover condition is satisfied, the method according to the present invention is applied, while, otherwise, the data transmission occurs according to the normal CAN standard. Also advantageously, the enlargement of the data field and the adaptation of the interpretation of the content of the data length code likewise occur as a function of a, for example the same, switchover condition. This makes it possible to use devices according to the present invention both in standard CAN bus systems as well as in new bus systems according to the present invention with potentially greater data fields.
The satisfaction of the switchover conditions is communicated to the recipients by one or multiple identifiers. For example, according to an advantageous example embodiment, at least one of the identifications occurs by a first identification bit, the position of which is between the last bit of the identifier and the first bit of the data length code and at the position of which, in messages according to the CAN standard ISO 11898-1, there is a bit having a defined value.
It is furthermore advantageous that possibly existing stuff bits, which appear before the CRC field in the message, are also included in the calculation of the checksum. This further improves the reliability of the data transmission or the detection probability for data transmission errors.
If the method is further combined with a switchover of the bit length, for example for at least the bits of the data field and the CRC field, then a further advantage is obtained that a greater quantity of data is transmitted in accelerated fashion than is the case when the data field is limited to 8 bytes. This further increases the average data transmission rate of the bus system. In one advantageous development, the messages of a shortened bit length are in this case identified by another identification bit, which is located between the first identification bit and the first bit of the data length code. This allows the switchover of the bit length to occur independently of the switchover of the CRC calculation or the size of the data field, and it is possible to react flexibly to the prevailing conditions of the bus system.
The method is advantageously applicable in the normal operation of a motor vehicle for transmitting data between at least two control units of the motor vehicle, which are connected via a suitable data bus. It may equally be used advantageously during the manufacturing or maintenance of a motor vehicle for transmitting data between a programming unit connected to a suitable data bus for programming purposes and at least one control unit of the motor vehicle that is connected to the data bus. It is also advantageously usable in the industrial field when larger data quantities must be transmitted for example for control purposes. Particularly if a reduced data rate must be applied during the arbitration due to the length of the transmission route so that all subscribers have the opportunity to access the bus, the method makes it possible, in particular in combination with the switchover of the length of the data field and the reduction of the bit length, to achieve a higher data transmission rate.
An additional advantage is that a standard CAN controller only needs to be modified minimally in order to be able to operate in accordance with the present invention. A communications controller according to the present invention, which is also able to work as a standard CAN controller, is only negligibly larger than a conventional standard CAN controller. The associated application program does not need to be modified, and even then advantages in the speed of data transmission are already achieved.
Advantageously, substantial portions of the CAN conformance test (ISO 16845) may be adopted. In one advantageous development, the transmission method according to the present invention may be combined with the supplements of TTCAN (ISO 11898-4).
In the following, the present invention and its advantages will be described with reference to drawings and example embodiments. The subject matter of the present invention is not limited to the represented and described example embodiments.
a shows the two alternatives for the structure of messages in the CAN format according to the CAN standard ISO 11898-1 from the related art.
b shows analogous alternatives for the format of the respective messages of
a shows the structure of messages as they are used in a CAN bus for data transmission. The two different formats “standard” and “extended” are shown. The method according to the present invention is applicable to both formats.
The message begins with a “start of frame” (SOF) bit, which signals the beginning of the message. This is followed by a section that is used primarily for identifying the message and on the basis of which the subscribers of the bus system decide whether they receive the message or not. This section is called an “arbitration field” and contains the identifier. This is followed by a “control field” containing, among other things, the data length code. The data length code contains information about the size of the data field of the message. This is followed by the actual “data field,” which contains the data to be exchanged between the subscribers of the bus system. This is followed by the “CRC field” including the 15-bit checksum and a delimiter, and subsequently two “acknowledge” (ACK) bits, which signal to the sender the successful reception of a message. The message is concluded by an “end of frame” (EOF) sequence.
In the standard CAN transmission method, the data field may contain a maximum of 9 bytes, that is, 64 bits of data. According to the standard, the data length code comprises four bits, which is to say that it can assume 16 distinct values. Of this value range, in today's bus systems, only eight different values are used for the various sizes of the data field from 1 byte to 8 bytes. A data field of 0 bytes is not recommended in standard CAN, and sizes above 8 bytes are not permitted. The assignment of the values of the data length code to the sizes of the data field is shown in
In
In the transmission method modified in accordance with the present invention, the data field may contain more than 8 bytes, namely, in the represented embodiment, up to K bytes. In contrast to standard CAN, additional values assumable by the data length code are used to identify larger data fields. For example, the four bits of the data length code may be used to represent the values from zero to 15 bytes. Other assignments may also be made, however, one possibility being the use the value of the data length code DLC=0b0000, which is normally not used in today's CAN messages, for another possible size of the data field, for the size of 16 bytes for example. These two possibilities are shown in
Another possibility is that, for the values of the data length code greater than 0b1000 and up to 0b1111, the associated sizes of the data field grow respectively by 2 bytes for example. This case is shown in the table as DLC 3. The maximum size of data field K in this variant reaches the value 24. By selecting a greater increment, for example 4 bytes, greater data fields would be achievable accordingly.
In the DLC 3 example, an additional modification has been made: the value DLC=0b0000 is used in this example embodiment by remote frames. Standard CAN, by contrast, provides for transmitting a remote frame at the same value of the DLC as the message transmitted as a reaction to the remote frame. The modification described here ensures that remote frames having a different DLC and an identical identifier cannot be transmitted, which (cf. ISO 11898-1, chap. 10.8.8) may result in inextricable collisions.
In the embodiments of the method shown in tabular form in
To ensure that such a communications controller is able to determine in what manner it must interpret the contents of the data length code, it is advantageous for the communications controller to recognize independently whether the communication of the bus system occurs according to standard CAN or according to the method of the present invention. According to an example embodiment, this is achieved by using a reserved bit within the arbitration field or the control field for identification such that, from this first identification K1, the communications controller is able to derive a first switchover condition UB1, on the basis of which it selects the transmission method. For example, the second bit of the control field indicated by r0 in
According to an alternative example embodiment, the determination is selected as a function of the identifier format. For standard addressing, one option for identifying the messages according to the present invention is thus to insert a recessive EDL (extended data length) bit into the control field in the position of the r0 bit that is always dominant in standard CAN. For extended addressing, the recessive EDL bit in the control field can take the position of the r1 bit that is always dominant in standard CAN.
An alternative example embodiment uses the SRR bit, which in standard CAN must always be transmitted recessively, but which is accepted also dominantly by the bus subscribers receiving the message. It is also possible to evaluate bit combinations to determine the first switchover condition UB1.
According to an alternative example embodiment, the use of the extended format is prescribed for the transmission method modified in accordance with the present invention. Messages in the extended format are recognized by bus subscribers on the basis of the value of the IDE bit (cf.
It is also possible to apply the method in suitable communications controllers that are not also designed for standard-conforming CAN communication. In this case, the determination of the mentioned first switchover condition UB1, for example as a function of a suitable identification K1 of the messages, may also be dropped. In this case, the communications controllers rather operate exclusively according to one of the described methods and are accordingly usable only in bus systems in which such communications controllers according to the present invention are used exclusively.
If, as provided in the present invention, the data field of messages is enlarged, then, according to an example embodiment, the method utilized for the cyclic redundancy check (CRC) is adapted in order to obtain a sufficient immunity against error. In particular, it may be advantageous to use a different CRC polynomial, for example of a higher order, and accordingly to provide a CRC field of a deviating size in the messages modified in accordance with the present invention. This is indicated in
The use of a modified method for calculating the CRC checksum may be signaled to the bus subscribers by a third identification K3, which represents a third switchover condition UB3. This identification K3 and the third switchover condition UB3, however, may also agree with the first identification K1 and/or switchover condition UB1. Here too, as was described further above, the reserved bit r0 from
In standard CAN controllers, the CRC code of CAN messages to be transmitted is generated by a feedback shift register, the serially transmitted bits of the message being fed sequentially into its input. The width of the shift register corresponds to the order of the CRC polynomial. The CRC encoding occurs by combining the register content with the CRC polynomial during the shift operations. When CAN messages are received, the serially received bits of the message are accordingly shifted into the CRC shift register. The CRC test is successful if at the end of the CRC field all bits of the shift register are at zero. The CRC code generation in the sending case and the CRC test in the receiving case both occur in hardware without requiring an intervention of the software. A modification of the CRC encoding thus does not affect the application software.
In the standard CAN protocol, the stuff bits within the CAN messages (cf. ISO 11898-1, chap. 10.5) are not included in the calculation or checking of the CRC code (cf. ISO 11898-1, chap. 10.4.2.6: “ . . . the bit stream given by the destuffed bit sequence . . . ”). This has the consequence that in rare cases two bit errors in one message are not detected even though the CRC should as such detect up to five randomly distributed bit errors in one message. This may occur when, as a result of the bit errors, stuff bits transform into data bit and vice versa (cf. Unruh, Mathony and Kaiser, “Error Detection Analysis of Automotive Communication Protocols,” SAE International Congress, No. 900699, Detroit, USA, 1990).
In the transmission method modified in accordance with the present invention, by contrast, the CRC encoding may be changed in such a way that the stuff bits within the message are also included in the calculation or checking of the CRC codes. That is, in this specific embodiment, the stuff bits belonging to the arbitration field, control field and data field are treated as part of the data to be protected by the cyclic redundancy check. As in standard CAN, the stuff bits of the CRC field are disregarded.
In an example embodiment, the communications controller is designed in such a way that it is compatible with the standard CAN, that is, it works in a standard-conforming fashion in a standard CAN bus system, while, in a bus system modified in accordance with the present invention, it allows for larger data fields in the messages and also performs the adapted calculation and checking of the CRC code.
Since, at the start of the reception of a message, it is not yet clear whether a standard-conforming CAN message or a message modified in accordance with the present invention is received, two CRC shift registers are implemented in a communications controller according to the present invention, which shift registers work in parallel. Following the reception of the CRC delimiter, when the CRC code is evaluated in the receiver, it is clear from the third identification K3 according to the present invention or from the third switchover condition UB3 derived from the identification or the content of the data length code, for example, which transmission method was used, and the shift register associated with this transmission method is then evaluated. As already explained above, the third switchover condition UB3 may agree with the first switchover condition UB1, which concerns the size of the data field and the interpretation of the data length code.
To be sure, it is already clear for the sender at the beginning of sending a message according to which transmission method a transmission is to occur. Since it could happen, however, that the arbitration regarding bus access is lost and the started message is not sent, but instead a different message is received, both CRC shift registers are activated in parallel in this case as well.
The described implementation of two CRC shift registers working in parallel also allows for another improvement, as follows. The CRC polynomial of the standard CAN protocol (x15+x14+x10+x8+x7+x4+x3+1) is designed for a message length of less than 127 bits. If messages transmitted in accordance with the present invention also use longer data fields, then it is practical to use a different, in particular longer, CRC polynomial in order to maintain transmission reliability. The messages transmitted in accordance with the present invention accordingly receive a modified, in particular longer, CRC field. In ongoing operation, the communications controllers switch dynamically between the two CRC shift registers, that is, between the standard CAN-conforming shift register and the shift register of the present invention, in order to use the respectively fitting polynomial.
Of course, more than two shift registers and accordingly more than two CRC polynomials may also be used, graduated as a function of the length of the data field or the desired transmission reliability. In this case, if a compatibility with the standard CAN is to be maintained, the corresponding identification and the associated switchover condition must be adapted. For example, a first switchover condition UB1 could be triggered by the reserved r0 bit or the SRR bit in
It is also possible that first switchover condition UB1 switches over to the option of longer data fields and the corresponding interpretation of the content of the data length code, perhaps via the reserved bit r0 or the SRR bit, and that the ascertainment of the third switchover condition UB3 and accordingly the selection of the CRC polynomial to be evaluated for the CRC check then occurs as a function of the content of the data length code. Third switchover condition UB3 may accordingly also assume more than two values. For example, the data field sizes could be selected according to DLC 3, that is, assume the values 0 (for remote frames) 1, . . . , 8, 10, 12, 14, 16, 18, 20 and 24 bytes, and three CRC polynomials could then be calculated in parallel via suitable shift registers, for example the standard CRC polynomial for data fields up to 8 bytes, a second CRC polynomial for data fields up to 16 bytes and a third CRC polynomial for data fields up to 24 bytes.
The subscriber station is first in a bus-scanning state as long as there is no communications traffic on the bus. Query 302 is thus waiting for a dominant bit on the bus. This bit marks the start of a new message.
As soon as the start of a new message has been determined, the calculation of the at least two checksums to be calculated in parallel begins in Block 304. The first checksum corresponds to the CRC calculation of the standard CAN, while the second checksum is calculated according to the new method. In the calculation of the second checksum, the stuff bits are included in the example embodiment shown, whereas this is not the case in the calculation according to the standard CAN. It is also possible, however, not to take the stuff bits into account even for calculating the second checksum, similar to the standard CAN.
Subsequently, beginning with step 306, the additional bits of the message following the SOF bit are received, beginning with the arbitration field. If multiple bus subscribers want to send out a message, then the bus subscribers negotiate among themselves in accordance with the usual method of the standard CAN which bus subscriber gains the access to the bus. Block 306 indicates the reception of all bits until first identification K1 has been received or first switchover condition UB1 has been determined. In the examples listed, first switchover condition UB1 is ascertained from the arbitration field, for example from the SRR bit or the IDE bit, or from the control field, for example from a reserved bit of the same (cf.
If it is known at branching 310, for example following the reception of the first two bits of the control field, that, according to first switchover condition UB1, the communication occurs in accordance with the CAN standard (the path of
If, by contrast, it is known at branching 310, for example following the reception of the first two bits of the control field, that, according to first switchover condition UB1, the communication method modified in accordance with the present invention is to be applied (the path of
In the reception process modified in this manner, as shown in
The transition between the states of Fast CAN Arbitration and Fast CAN Data may occur as a function of a fourth switchover condition UB4, which corresponds to a fourth identification K4 of the messages, which signals to the subscribers of the data transmission that the shortened bit length is used. In the example embodiment shown here, the chosen position of this identification K4 is the “reserved bit” r0, which is transmitted before the data length code. It thus corresponds to a possible position of first identification K1, which corresponds to first switchover condition UB1 and indicates the possible use of longer data fields and a modified interpretation of the data length code, and also of third identification K3, which corresponds to a modified CRC calculation.
Another possibility for identifying the messages having a shortened bit length in accordance with the present invention is shown in
The messages bear the label “CAN FD Fast.” For the two possible addressing variants of messages, the standard format and the extended format, areas are drawn in in
In the case shown, in which the first identification EDL is thus followed by the second identification BRS, messages are transmitted in the transmission method according to the present invention, the bit length of the messages is markedly shortened, the data field size of the messages is expandable to values above 8 bytes, and the CRC of the messages is adapted to the larger data field. A substantial increase of the transmission capacity via the bus system is thus achieved while the transmission reliability is at the same time improved.
In the example shown, the faster transmission begins immediately after the transmission of the associated identification and is ended immediately after reaching the bit defined for the reverse switchover or when a reason for starting an error frame was detected.
The following calculation illustrates the utility in terms of the achieved data transmission rate of the example embodiment shown in
Under the given boundary conditions, an effective transmission performance of 293 bits in 160 microseconds is achieved, which, at an identical assumed bus capacity utilization, corresponds to a data transmission rate that is increased by a factor of 3.7 compared to unmodified standard CAN transmission. Additionally, the ratio of useful data (data field) to protocol overhead shifts advantageously.
The method is suitable in the normal operation of a motor vehicle for transmitting data between at least two control units of the motor vehicle, which are connected via a suitable data bus. It may also be used advantageously during the manufacturing or maintenance of a motor vehicle for transmitting data between a programming unit connected to a suitable data bus for programming purposes and at least one control unit of the motor vehicle that is connected to the data bus.
It is furthermore also possible to use the method in industrial automation, that is, for example in the transmission of control information between distributed control units interconnected by the bus, which control the course of an industrial manufacturing process. In this environment there may also be very long bus lines and it may be particularly practical to operate the bus system for the arbitration phase at a relatively long bit length, for example at 16, 32 or 64 microseconds, so that the bus signals, as required, are able to propagate through the entire bus system during the arbitration process. Subsequently, a switchover to shorter bit lengths may be performed for part of the message, as described, so as not to let the average transmission rate drop too low.
Overall, the method represents a transmission method that is characterized by the fact that a standard CAN controller only needs to be modified minimally in order to work in accordance with the present invention. A communications controller according to example embodiments the present invention, which is also able to work as a standard CAN controller, is only negligibly larger than a conventional standard CAN controller. The associated application program does not need to be modified, and even then advantages in the speed of data transmission are already achieved. Using the extended size of the data field and the associated DLC and CRC makes it possible to increase the speed of data transmission further, while adaptations in the application software are minimal. Advantageously, substantial portions of the CAN conformance test (ISO 16845) may be adopted. It is also possible to combine the transmission method according to the present invention with the supplements of TTCAN (ISO 11898-4).
Number | Date | Country | Kind |
---|---|---|---|
10 2011 006 875.9 | Apr 2011 | DE | national |
10 2011 078 266.4 | Jun 2011 | DE | national |
10 2011 080 476.5 | Aug 2011 | DE | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/EP2012/055574 | 3/29/2012 | WO | 00 | 3/19/2014 |