CN101582799B - Bus switch method based on CAN redundance - Google Patents

Abstract

The invention discloses a bus switching method based on CAN redundancy, which comprises the following steps: 1), judging whether it is a sending node or a receiving node; if it is a sending node, enter step 2); if it is a receiving node, enter step 5); 2) , judging whether the sending node sends overtime; judged as "yes", enter step 3), judged as "no", then enter step 4); 3), switch bus to resend; 4), send node monitor bus until timeout or receive To the response frame of the receiving node; 5), judge whether the receiving node has received a timeout; if it is judged as "yes", enter step 6), if it is judged as "no", then enter step 7); 6), switch the bus by broadcasting, inform other The node switches the bus to resend the data; 7), the receiving node sends a response frame as a response to the sending node, indicating that the data is sent and received successfully. The method can guarantee reliable correctness and real-time performance of data reception.

Description

Bus Switching Method Based on CAN Redundancy

technical field

The invention belongs to the technical field of CAN network communication, in particular to a bus switching method based on CAN redundancy.

Background technique

An effective way to solve CAN network reliability communication is to carry out different degrees of redundancy on the bus. Although CAN redundancy is a common method to solve the reliability of CAN network, no one has studied the bus switching method based on CAN redundancy. Many designers switch the bus when sending or receiving fails to cause a timeout. Undoubtedly, during the period of waiting for the timeout, the node receiving the wrong data will miss a lot of valid data frames, which reduces the bus utilization. However, a good bus switching method based on CAN redundancy should be able to make the receiving node miss data frames as little as possible, thereby improving the bus utilization rate, ensuring the correct and stable transmission of data in the CAN network, and enhancing the reliability of the CAN network. real-time.

Contents of the invention

The technical problem to be solved by the present invention is to provide a bus switching method based on CAN redundancy that can ensure reliable correctness and real-time performance of data reception.

In order to solve the problems of the technologies described above, the present invention provides a bus switching method based on CAN redundancy, comprising the following steps:

1), determine whether it is a sending node or a receiving node; when it is a sending node, enter step 2); when it is a receiving node, enter step 5);

2), judge whether the sending node sends overtime; when it is judged as "yes", enter step 3), when judged as "no", then enter step 4);

3), switch the bus and resend;

4), the sending node monitors the bus until it times out or receives a response frame from the receiving node, indicating that the data is successfully sent;

5), determine whether the receiving node receives timeout; when it is judged as "yes", enter step 6), when it is judged as "no", then enter step 7),

6) Switch the bus by broadcasting to inform other nodes to switch the bus and resend the data;

7) The receiving node sends a response frame as a response to the sending node, indicating that the data is sent and received successfully.

As the improvement of the bus switching method based on CAN redundancy of the present invention: in step 2), the criterion for judging whether the sending node sends overtime is: whether the sending time of the sending node exceeds the longest sending time set by the user, exceeding this value is considered The data sending failed this time.

As a further improvement of the bus switching method based on CAN redundancy of the present invention: in step 5), the criterion for judging whether the receiving node sends overtime is: whether the receiving error of the receiving node exceeds a set threshold, which is user-defined and can be tolerated The maximum number of receiving errors.

As a further improvement of the bus switching method based on CAN redundancy in the present invention: only one bus can send data at the same time; use the software buffer as the buffer pool for the user to send data, and use the three hardware buffers of the development board to accelerate data frames The transmission; set the threshold and timeout mechanism to restart the controller or switch the bus to resend or continue to receive data frames when the communication fails; use the node indicator to remind the user that the bus is faulty, and the bus needs to be restored in time.

As a further improvement of the bus switching method based on CAN redundancy of the present invention: step 3) is specifically to carry out the following steps in sequence:

A), restart the controller and resend; if the sending is successful, go to step B); if the sending is still unsuccessful, so that the sending times out, go to step C);

B), monitor and wait for the response frame on the network, when receiving the response frame of the receiving node, it means that the data transmission is successful;

C), judge whether all the other buses are available, if any one (at least one) of the other buses is available, then use this bus to resend; if all the other buses are unusable, then prompt the user and wait for the user to recover.

As a further improvement of the bus switching method based on CAN redundancy of the present invention: step 6) is specifically to carry out the following steps in sequence:

Determine whether the rest of the buses are available. If any (at least one) of the rest of the buses is available, then broadcast to switch the bus and restart the original controller (indicating that the node cannot receive valid data frames on the bus); if another If the bus is not available, restart the original controller.

The above content is shown in Figure 1: when the bus is switched, another set of controllers and transceivers are used. Before the start of network transmission, the controller has been initialized, but because some registers of the controller have to be initialized, or some software counts errors, so the controller needs to be re-initialized, which means restarting the controller.

In the bus switching method based on CAN redundancy in the present invention: a software buffer structure array (Buffer) for recording and sending data frames is designed at each node, so that the data frames do not need to be transferred when the bus is switched. The condition for judging whether the software buffer is full is: whether the valid data in the data buffer==NUM-1. NUM is the maximum size of the software buffer. The reason for subtracting 1 is to reserve a data buffer for the special data frame; the reason for reserving a data buffer for the special frame is to prevent the special frame from being unable to be sent in time when the software buffer is full of data frames. It must be ensured that special data frames can be sent out in time. When sending a special frame, insert it into the reserved position of the software buffer to ensure that it is sent first. The special data frame includes the bus switching frame Change_Send_Frame of the sending node and the bus switching frame Change_REV_Frame of the receiving node. Change_Send_Frame refers to the switching bus frame sent to each node when the sending node times out, notifying each node that the original channel has been broken; and Change_REV_Frame is the switching bus frame broadcast by the receiving node when the receiving error exceeds the threshold, which does not mean that the original bus appears Faults are just switching buses to minimize the loss of data frames at the receiving node.

When CAN redundancy is used to improve the reliability of CAN network communication, the CAN redundancy-based bus switching method of the present invention is beneficial to improving bus utilization, network communication reliability and security, and network real-time performance. The bus switching method based on CAN redundancy of the present invention is used to ensure that the sending node can successfully send data, the receiving node can receive the data, and the data frame is lost as little as possible, thereby ensuring the reliability, correctness, real-time performance and accuracy of data reception. safety.

In the present invention, it is possible to avoid waiting until receiving overtime due to bus hardware failure (in a CAN network, the overtime period setting must be longer than the time when the error data exceeds the threshold), and during this period of waiting, other nodes may be missed Data transferred on the faulty bus. To put it simply, the receiving node does not set a timeout period, and switches the bus to send data when the receiving error exceeds the threshold. Note: (1) If the receiving error count exceeds the threshold, the bus will be switched, but it does not mean that the bus is broken, but to minimize the loss of data frames, so that the node can receive data frames in time. (2) When the sending error count exceeds the threshold, the controller will be restarted without switching the bus. After all, switching the bus is more expensive than restarting the controller. When the transmission is still not successful after restarting, a timeout interrupt will occur to switch the bus and resend.

Taking the dual CAN network as an example, the present invention sets a threshold and a timeout mechanism for restarting the CAN controller or switching the bus when a communication failure occurs, which is mainly reflected in:

●When the sending node sending error count value exceeds a certain threshold, restart the CAN controller to resend; if restarting the controller still fails to send, so that when the sending timeout occurs, then: if another bus is available, switch the bus and resend; if If the other bus is unavailable, both buses are disabled, and the user is notified (waiting for user recovery).

●When the receiving node receives an error count value exceeding a certain threshold: If another bus is available, restart the CAN controller that the node is using, broadcast to switch the bus, (broadcast and send special data frames on the CAN network, and notify each node: The node cannot receive valid data frames on the bus, and all nodes need to switch to another bus to send data); if another bus is unavailable, restart the CAN controller of the node.

The sending timer of the sending node is timed by the system clock. It starts timing when the data frame is put into the software buffer, and assigns the member variable time value of the data frame to the current system clock tick value plus the threshold cycle tick number. . Every tick++ will judge whether the data frame time at the head of the software buffer has timed out. The number of sending errors and receiving errors is automatically recorded inside the CAN controller. By setting the relevant registers of the CAN controller, the transmitter and receiver trigger an interrupt when the error count exceeds a certain threshold. If the receiving error count value exceeds the threshold, the bus will be switched, but it does not mean that the other bus is broken, but to minimize the loss of data frames, it is not necessary to wait until the timeout before switching the bus. When the transmit error count exceeds the threshold, the controller is restarted without switching the bus. Because switching the bus is more expensive than restarting the controller. When the transmission is still not successful after restarting, a timeout interrupt will occur to switch the bus and resend.

Each node designs a software buffer structure array (Buffer) for recording data frames, so that the data frames do not need to be transferred when the bus is switched. The data structure of the software buffer data frame Soft_Frame (that is, the data structure of the array elements) is as follows:

typedef struct Soft_Frame{INT32U id;//The bus switching frame sent by the sending node, the bus switching frame sent by the receiving node //The IDs of the bus recovery frame are 0, 1, 2 INT8U type; //The type of data frame (extended frame, Standard frame, data frame, remote frame,) INT8U data[8]; INT8U datalen; INT8U prio;//buffer priority (that is, data frame priority)

The condition for judging whether the software buffer is full is: whether the number of valid data in the buffer is equal to NUM-1 (NUM is the maximum capacity of the software buffer). In order to prevent the special frame from being unable to be sent when the software buffer is full of data frames, we need to reserve a data buffer in the software buffer for the special frame. When sending a switch bus frame, we insert it into the reserved position of the software buffer and send it first. Therefore, when the number of data frames in the software buffer is NUM-1, it means that the software buffer is full.

We assign priority to normal data frames incrementally in the order they arrive in the software buffer. In order to ensure that the data frames are sent according to the principle of first-come-first-send, when the priority increases to 255 (the priority register in the hardware buffer is 8 bits, there are at most 256 different priorities), the user continues to the software buffer Putting in a data frame will return an error, and the software buffer will not be able to receive the user's data frame until the data frame with a priority of 255 is sent, and the priority of the data frame will increase from 1.

The special data frames include the bus switching frame Change_Send_Frame of the sending node, the bus switching frame Change_REV_Frame of the receiving node, and the bus recovery frame Recov_Frame. The bus switching frame of the sending node refers to the switching bus frame sent to each node when the sending node times out, notifying each node that the original channel has been broken; the bus switching frame of the receiving node is the switching broadcasted by the receiving node when the receiving error exceeds the threshold The bus frame does not mean that the original bus is faulty, but it is just a measure to switch the bus to minimize the loss of data frames of the receiving node; the bus recovery frame is the user's notification to each node that the original broken channel can be restored. Its IDs are 0, 1, 2 respectively; and the priority of Change_REV_Frame and Change_Send_Frame is 0 (these two data frames will not appear at the same time, so the priority can be the same), which means that when this type of frame reaches the reserved software buffer position, it will be the first data frame to be successfully sent (because the sending of other data frames in the current software buffer is blocked on another bus); the priority of other frames must not be 0; the priority of Recov_Frame It is the same as the general data frame, otherwise it will affect the operation of removing the first data frame of the software buffer for sending successful interrupts, because sending successful interrupts will always delete the reserved position or the first position of the software buffer (if the reserved position is empty) , Recov_Frame arrives at the software buffer during the normal transmission of other data frames. If it is placed in the position of a special frame, it will be successfully sent and the ISR will mistakenly delete the special frame, but in fact the data frame has not been sent yet. The best solution is to place the bus recovery frame at the fourth frame position, but the overhead is too large, and all data frames after the fifth position need to be moved backwards. This design does the following processing: switch the bus frame directly into the chain head of the software buffer to ensure that it is sent first; the bus recovery frame is treated as a general data frame and inserted into the end of the software buffer array.

The experiment proves that: when the bus in use fails, the design can send data frames normally, and the receiving node using the bus switching method loses less data frames than the nodes not using the bus switching method. The method can ensure that the sending node can successfully send data, the receiving node can correctly receive data, and the data frame is lost as little as possible, thereby ensuring the reliability, correctness and real-time performance of data reception.

Description of drawings

The specific implementation manners of the present invention will be described in further detail below in conjunction with the accompanying drawings.

Figure 1 is a CAN redundancy structure diagram;

Fig. 2 is a flow chart when Fig. 1 sends data;

Fig. 3 is the flow chart when Fig. 1 receives data;

Fig. 4 is a flowchart when Fig. 1 resends data;

Fig. 5 is the interrupt flowchart when Fig. 1 data is sent successfully;

Fig. 6 is a flowchart of Fig. 1 error exceeding threshold processing;

FIG. 7 is a flow chart of bus switching over time in FIG. 1 .

Detailed ways

Embodiment 1, a double CAN network as shown in Figure 1, adopts two sets of buses (i.e. bus A and bus B), and each set of buses includes complete bus cables, bus drivers, and bus controllers.

In order to realize the bus switching method based on CAN redundancy of the present invention in above-mentioned double CAN network, therefore need to carry out following processing:

When the sending node sending error count value exceeds a certain threshold, restart the CAN controller to resend. Restarting the CAN controller is mainly to clear some register errors in the controller; Then: if another bus is available, switch the bus and resend; if the other bus is not available, both buses will be disabled, and the node indicator light will be used to prompt the user, waiting for the user to detect the bus path of the node and recover in time.

When the receiving node receives an error count value exceeding a certain threshold: if another bus is available, restart the CAN controller that the node is using, broadcast the switching bus (broadcast and send special data frames on the CAN network, and notify each node: the node Valid data frames cannot be received on this bus, and all nodes need to switch to another bus to send data); if another bus is unavailable, restart the CAN controller of the node until it times out, the node stops and waits User recovery.

Embodiment 2, the bus switching method based on CAN redundancy that is carried out to the dual CAN network described in embodiment 1, two CAN buses are opened simultaneously, but at most only one bus sends data at a time, in order to avoid problems caused by asynchronous Received data error. Use the software buffer as the buffer pool for the user to send data. When the three hardware buffers of the node are free, put the data in the software buffer into the hardware buffer to send the data frame; set the threshold and timeout mechanism It is used to restart the CAN controller or switch the bus when there is a communication failure; use the on and off of the LED lights of each bus node to prompt the user for bus failure and restore the bus in time.

Specifically, follow the steps below:

1), determine whether the node belongs to the sending node or the receiving node, if it belongs to the sending node, enter step 2), if it belongs to the receiving node, enter step 7);

2), the flow of the sending node sending data is shown in Figure 2: record the data sending timeout point deadline at the beginning of the data frame sending command, then put the data frame into the software buffer, if the hardware buffer is free, then further Put it into the hardware buffer and wait for sending, and the hardware register will automatically record the number of sending errors during the sending process. When the transmission error exceeds the threshold set by the user, it will enter step 3); otherwise, when the response frame from the receiving node is received, the transmission is successful, and it will enter step 4).

3) The interrupt processing program of the error count exceeding the threshold, as shown in Figure 6: When the sending error exceeds the threshold, it is judged whether the sending node has timed out, that is, it is judged whether the sending time of the sending node exceeds the longest sending time set by the user, If the value exceeds the value, go to step 5), otherwise restart the controller, resend the data, and go to step 6); when the receiving error exceeds the threshold, it is judged whether another bus is available. If another bus is available, the broadcast switches to available bus, send and receive data, and restart the controller of the original bus; if the other one is not available, restart the original controller and prompt the user (for example, in the form of an indicator light), waiting for the user to recover, the user is on the first CAN bus When deactivated, you should find a way to restore the bus in time.

4) Successfully sent data interrupt processing is specifically shown in Figure 5: remove the data frame sent in the software buffer, and read the software buffer, if the data frame to be sent can be read, put it into the interrupt to generate the The hardware buffer is waiting to be sent, and the hardware buffer empty flag is cleared; if no data frame is read, it means that the software buffer is empty, and there is no data frame waiting to be sent, and the hardware buffer empty interrupt is disabled.

5) The processing flow of bus switching is shown in Figure 7: if another bus is available, re-initialize the original controller, set the indicator light to remind the user that the original bus is down, wait for the user to recover in time, and then switch to the available bus to resend data frame, that is, enter step 6); if another bus is unavailable, reinitialize the controller and set the channel to hang up, light on the indicator to prompt the user node may hang up, and recover in time.

6) The process of resending data is shown in Figure 4: when the data size of the software buffer is 0, an error is returned; when the data is not 0, if there is a special frame, start timing the special frame and put the special frame first If there is no special frame, put the data frame at the buffer+head position into the hardware buffer and wait to be sent.

7) The receiving process of the receiving node is shown in Figure 3: when the receiving error count does not exceed the threshold, the data has been successfully received, and the data is considered to be received successfully, and a response frame will be automatically sent to the sender. The receiving node processes data or bus status according to the type of data frame; if the receiving error exceeds the threshold, it turns to step 3).

In order to prove that the method of the present invention can indeed guarantee reliable correctness and real-time performance of data reception, the inventor has done the following comparative experiments:

The CAN redundant network described in Embodiment 1 is selected for use. Then according to the method described in embodiment 2, according to the fault that CAN communication network system may cause (such as the disconnection or short circuit of cable, the fault of CAN bus driver, the fault of bus controller or even single-chip microcomputer, the interference of CAN network external magnetic field etc. ), test and compare the network that has adopted the bus switching method that the present invention proposes, and the network that does not adopt this method, the difference on the stability of data frame reception, correctness, real-time performance.

The hardware environment tested by this method:

1. Three HCS12 development boards: used as CAN nodes (the HCS12 development board integrates the CAN controller into the development board).

2. Several CAN buses: used to connect to a CAN network.

3. Programmer: used to program the application program to the HCS12 development board.

The software environment tested by this method:

1. CAN debugging tool ZLGCANTest: used to test the data frames received by CAN.

2.CodeWarrior IDE: used to compile the code programmed into HCS12.

The test can prove that the method of the present invention can enable the sending node to successfully send data, the receiving node to receive the data, and lose data frames as little as possible, thereby ensuring the reliability, correctness and real-time performance of data reception.

Finally, it should be noted that the above examples are only some specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments, and many variations are possible. All deformations that can be directly derived or associated by those skilled in the art from the content disclosed in the present invention should be considered as the protection scope of the present invention.

Claims (6)

1. A bus switching method based on CAN redundancy, characterized in comprising the following steps: 1), determine whether it is a sending node or a receiving node; when it is a sending node, enter step 2); when it is a receiving node, enter step 5); 2), judge whether the sending node sends overtime; when it is judged as "yes", enter step 3), when judged as "no", then enter step 4); 3), switch the bus and resend, and the process ends; 4) The sending node monitors the bus until it times out or receives a response frame from the receiving node, indicating that the data is successfully sent and the process ends; 5), determine whether the receiving node receives timeout; when it is judged as "yes", enter step 6), when it is judged as "no", then enter step 7); 6) Switch the bus by broadcasting, inform other nodes to switch the bus and resend the data, and the process ends; 7) The receiving node sends a response frame as a response to the sending node, indicating that the data is sent and received successfully, and the process ends. 2. a kind of bus switching method based on CAN redundancy according to claim 1 is characterized in that: the standard of judging whether sending node sends overtime in described step 2) is: whether the sending time of sending node exceeds user setting The longest sending time, if it exceeds this value, it will be considered that the data sending failed. 3. a kind of bus switching method based on CAN redundancy according to claim 1, it is characterized in that: in the described step 5), the standard of judging whether the receiving node sends overtime is: whether the receiving error of the receiving node exceeds the set Threshold, which is the user-defined maximum number of receiving errors that can be tolerated. 4. according to claim 1,2 or 3 described a kind of bus switching method based on CAN redundancy, it is characterized in that: there can only be a bus to send data at the same moment; Use software buffer as the buffer pool of user sending data, Use the three hardware buffers of the development board to speed up the sending of data frames; set the threshold and timeout mechanism to restart the controller or switch the bus to resend or continue to receive data frames when the communication fails; Restore the bus in time. 5. a kind of bus switching method based on CAN redundancy according to claim 4, is characterized in that: described step 3) is specifically to carry out following steps successively: A), restart the controller and resend: if the sending is successful, go to step B); if the sending is still unsuccessful, so that the sending times out, go to step C); B), monitor and wait for the response frame on the network, when receiving the response frame from the receiving node, it means that the data transmission is successful, and the process ends; C), judging whether the remaining buses are available, if at least one of the remaining buses is available, then use this bus to resend; if the remaining buses are not available, then prompt the user, wait for the user to recover, and the process ends. 6. a kind of bus switching method based on CAN redundancy according to claim 5, is characterized in that: described step 6) is specifically to carry out following steps successively: Determine whether the remaining buses are available. If at least one of the remaining buses is available, broadcast the switching bus and restart the original controller, indicating that the node cannot receive valid data frames on this bus; if another bus is unavailable, restart The original controller, the process ends.

Images