WO2021199607A1 - Control device and method - Google Patents

Abstract

Provided is a control device and a method that can suppress reduction in processing accuracy. A first task unit executes a first task (cyclic task A) in a predetermined execution cycle and writes an output value of the first task in a buffer. A predicted task unit predicts the output value of the first task before the output value of the first task is written in the buffer in the execution cycle, and writes the predicted value in the buffer. A second task unit reads the output value or the predicted value of the first task from the buffer in the execution cycle, executes a second task (cyclic task B) using the output value or the predicted value read from the buffer as input, and outputs an output value of the second task.

Description

Control device and method

The present invention relates to a control device and a method.

The control device processes the input data and outputs the data to the subsequent device. For controls that require real-time performance, there are restrictions on data input and output timing, and the design is performed so as to satisfy the restrictions. The periodic task is processed within a predetermined time (from the start timing of the periodic task to the deadline). The deadline must be determined so that the system does not fail.

Patent Document 1 describes a technique for synchronizing tasks and communications. In this document, the time-determining operation of the succeeding task is guaranteed by holding the old output data of the task in the output memory area and transferring it to the input memory of the succeeding task.

Patent Document 2 describes a technique for monitoring the degree of delay of a task of a vehicle control device. In the same document, the delay degree of the subtasks that compose the task is monitored, and if the delay degree is equal to or higher than the threshold value, a fail-safe task is executed instead of the task, and the task cannot complete the process within the deadline and the entire system. Prevents it from shutting down.

Special Table 2016-523409 Gazette Japanese Unexamined Patent Publication No. 2016-66139

The deadline of a periodic task is generally determined based on the worst execution time calculated in advance. However, as the system becomes larger and more complex, it has become difficult to calculate the worst execution time. Further, due to the demand for high-speed processing, it is becoming difficult to adopt a deadline based on the worst execution time that occurs only with a very low probability. Therefore, it is necessary to set an approximately sufficient processing cycle and take measures in the unlikely event that the cycle cannot be observed.

If a certain periodic task did not complete the process within the specified time (overrun), the accuracy of the process had to be greatly reduced by using the previous value.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a control device and a method capable of suppressing a decrease in processing accuracy.

In order to achieve the above object, the control device which is an example of the present invention includes a buffer and a first task unit which executes the first task in a predetermined execution cycle and writes the output value of the first task to the buffer. , The prediction task unit that predicts the output value of the first task and writes the predicted value to the buffer before the output value of the first task is written to the buffer in the execution cycle, and the said in the execution cycle. A second task that reads the output value or the predicted value of the first task from the buffer, executes the second task with the output value or the predicted value read from the buffer as an input, and outputs the output value of the second task. It has two task units.

According to the present invention, it is possible to suppress a decrease in processing accuracy. Issues, configurations and effects other than those described above will be clarified by the description of the following embodiments.

It is a block diagram which shows an example of the structure of the vehicle control device which concerns on Example 1 of this invention. It is a figure which shows an example of the structure of the functional element of the vehicle control device which concerns on Example 1 of this invention. It is a flowchart which shows an example of the process performed in the periodic task A which concerns on Example 1 of this invention. It is a flowchart which shows an example of the process performed in the prediction task which concerns on Example 1 of this invention. It is a flowchart which shows an example of the process performed in the periodic task B which concerns on Example 1 of this invention. It is a flowchart which shows an example of the process performed by the scheduler which concerns on Example 1 of this invention. It is a figure which schematically explains the operation of each process in Example 1 of this invention on the time axis. It is a flowchart which shows an example of the process performed by the scheduler which concerns on Example 2 of this invention. It is a figure which schematically explains the operation of each process in Example 2 of this invention on the time axis. It is a flowchart which shows an example of the process performed in the prediction task which concerns on Example 3 of this invention. It is a figure which schematically explains the operation of each process in Example 3 of this invention on the time axis. It is a flowchart which shows an example of the process performed in the periodic task A which concerns on Example 4 of this invention. It is a flowchart which shows an example of the process performed by the scheduler which concerns on Example 4 of this invention. It is a flowchart which shows an example of the process performed in the periodic task A which concerns on Example 5 of this invention. It is a flowchart which shows an example of the process performed by the scheduler which concerns on Example 5 of this invention. It is a figure which schematically explains the operation of each process in Example 5 of this invention on the time axis. It is a block diagram which shows an example of the structure of the engine control unit which concerns on Example 6 of this invention. It is a figure which shows an example of the structure of the functional element of the engine control unit which concerns on Example 6 of this invention. It is a flowchart which shows an example of the process performed in the ignition timing calculation task which concerns on Example 6 of this invention. 6 is a flowchart showing an example of processing performed by the timer set task according to the sixth embodiment of the present invention. 6 is a flowchart showing an example of processing performed in the prediction task according to the sixth embodiment of the present invention. It is a block diagram which shows an example of the structure of the AD unit which concerns on Example 7 of this invention. It is a figure which shows an example of the structure of the functional element of the AD unit which concerns on Example 7 of this invention. It is a flowchart which shows an example of the process performed in the vertical position estimation task which concerns on Example 7 of this invention. It is a flowchart which shows an example of the process performed in the horizontal position estimation task which concerns on Example 7 of this invention. FIG. 7 is a flowchart showing an example of processing performed in the route calculation task according to the seventh embodiment of the present invention. It is a flowchart which shows an example of the process performed in the vertical position prediction task which concerns on Example 7 of this invention. It is a flowchart which shows an example of the process performed in the horizontal position prediction task which concerns on Example 7 of this invention.

Hereinafter, examples will be described with reference to the drawings. The present invention is not limited to the following examples, and the configuration of each part can be added, changed, or deleted as appropriate without departing from the spirit of the present invention. Here, as an example, a mode in which the present invention is applied to a vehicle control device is shown. The vehicle control device according to the following embodiment aims to prevent the accuracy from being lowered when the processing is overrun, for example.

[Example 1]
FIG. 1 is a block diagram showing an example of the configuration of the vehicle control device 1 according to the first embodiment. The vehicle control device 1 includes a processor 2, a memory 3, a non-volatile memory 4, and a buffer 15.
The processor 2 to the buffer 15 are connected via the interconnect 8.

The OS (Operating System) 7, the periodic task A (11), the prediction task 12, and the periodic task B (13) are loaded into the memory 3 and executed by the processor 2. In the following, for convenience, the task may be described as the operating subject of each process, but the actual operating subject of each process is the processor 2.

The periodic task A (11), the prediction task 12, the periodic task B (13), and the scheduler 14 are software that runs on the OS 7. As the OS7, for example, an OS compliant with AUTOSAR (Automotive Open System Archive) or a real-time OS can be used.

FIG. 2 is a configuration diagram of functional elements of the vehicle control device 1 according to the first embodiment.

In the vehicle control device 1, the periodic task A (11) performs processing and stores the output value in the buffer 15. The periodic task B (13) reads the value stored in the periodic task A (11) from the buffer 15 and performs processing. The prediction task 12 predicts the output value of the periodic task A (11) and stores it in the buffer 15. The scheduler 14 starts and ends the periodic task A (11), the prediction task 12, and the periodic task B (13).

In other words, the vehicle control device 1 (control device) includes at least a buffer 15, a first task unit, a prediction task unit, and a second task unit. In this embodiment, the first task unit, the prediction task unit, and the second task unit are realized by the cooperation of the software loaded in the memory 3 (FIG. 1) and the processor 2 (FIG. 1), respectively. However, it may be realized by hardware such as FPGA (Field-Programmable Gate Array).

Here, the first task unit (processor 2) executes the first task (cycle task A) in a predetermined execution cycle, and writes the output value of the first task (cycle task A) to the buffer 15. The prediction task unit (processor 2) predicts the output value of the first task (cycle task A) before the output value of the first task (cycle task A) is written to the buffer 15 in the execution cycle, and calculates the predicted value. Write to buffer 15. The second task unit (processor 2) reads the output value or predicted value of the first task (cycle task A) from the buffer 15 in the execution cycle, and inputs the output value or predicted value read from the buffer 15 to the second task (processor 2). Executes the periodic task B) and outputs the output value of the second task (periodic task B). Details of these operations will be described later with reference to FIG.

As a result, even if the output value of the first task (cycle task A) is not written to the buffer 15 in the execution cycle, the second task (cycle task B) is completed by inputting the predicted value of the first task (cycle task A). can do. As a result, it is possible to suppress a decrease in the accuracy of the output value (for example, the control amount of the actuator) of the second task (periodic task B).

The predicted value is written to the buffer 15, and then overwritten with the output value of the first task (periodic task A). A ring buffer may be used instead of overwriting.

FIG. 3 is a flowchart showing an example of the processing performed in the periodic task A (11) according to the first embodiment. The process 100 of the periodic task A (processor 2) is executed as follows.

In step 101, the periodic task A acquires the input value.

In step 102, the periodic task A performs arithmetic processing using the acquired input value.

In step 103, the periodic task A writes the result of the arithmetic processing as an output value in the buffer 15.

FIG. 4 is a flowchart showing an example of the processing performed in the prediction task 12 according to the first embodiment. The processing 200 of the prediction task (processor 2) is executed as follows.

In step 201, the prediction task acquires the input value.

In step 202, the prediction task predicts the value output in the periodic task A (11) using the acquired input value.

In step 203, the prediction task writes the obtained predicted value to the buffer 15.

FIG. 5 is a flowchart showing an example of the processing performed in the periodic task B (13) according to the first embodiment. The process 300 of the periodic task B (processor 2) is executed as follows.

In step 301, the periodic task B acquires the input value from the buffer 15.

In step 302, the periodic task B performs arithmetic processing using the acquired input value.

In step 303, the periodic task B outputs the result of the arithmetic processing as an output value.

FIG. 6 is a flowchart showing an example of processing performed by the scheduler 14 according to the first embodiment. The process 400 of the scheduler 14 (processor 2) is executed as follows.

In step 401, the scheduler 14 compares the time with the periodic task A start time 16. The scheduler 14 proceeds to step 402 if the time matches the periodic task A start time 16 (step 401: Yes), and if the time does not match the periodic task A start time 16 (step 401: No), the step Proceed to 403.

In step 402, since the time is the periodic task A activation time, the scheduler 14 activates the periodic task A (11).

In step 403, the scheduler 14 compares the time with the periodic task A deadline, and further determines whether the periodic task A (11) is being executed. When the time coincides with the periodic task A deadline and the periodic task A (11) is being executed (step 403: Yes), the scheduler 14 proceeds to step 404, and the time does not coincide with the periodic task A deadline. Alternatively, if the periodic task A (11) is not being executed (step 403: No), the process proceeds to step 405.

In step 404, the scheduler 14 ends the periodic task A (11).

In step 405, the scheduler 14 compares the time with the periodic task B start time. The scheduler 14 proceeds to step 406 if the time matches the periodic task B startup time (step 405: Yes), and if the time does not match the periodic task B startup time (step 405: No), step 407. Proceed to.

In step 406, since the time is the periodic task B activation time, the scheduler 14 activates the periodic task B (13).

In step 407, the scheduler 14 compares the time with the predicted task start time.
The scheduler 14 proceeds to step 408 if the time matches the predicted task start time (step 407: Yes), and proceeds to step 401 if the time does not match the predicted task start time (step 407: No). ..

In step 408, since the time is the predicted task activation time, the scheduler 14 activates the predicted task 12.

FIG. 7 is a diagram schematically explaining the operation of each process in the first embodiment on the time axis.

The black arrow in the figure represents the instruction of the scheduler 14, and the white arrow represents the reading from the buffer 15 or the writing to the buffer 15.

This example describes the case where the periodic task A (11-2) causes an overrun at the time (t) = 26.

The scheduler 14 activates the prediction task 12 at the prediction task start time 20 (step 408). As a result, the prediction task 12-1 is activated. When the prediction task 12-1 completes the calculation, the prediction task 12-1 writes the prediction value to the buffer 15. The processing time of the prediction task 12-1 is sufficiently smaller than the processing time of the periodic task A or the processing time of the periodic task B (the same applies to 12-2 described later).

The scheduler 14 activates the periodic task A (11) at the periodic task A start time 21. As a result, the periodic task A (11-1) is activated. When the calculation is completed, the periodic task A (11-1) writes an output value to the buffer 15.

The scheduler 14 activates the periodic task B (13) at the periodic task B start time 23. As a result, the periodic task B (13-1) is activated. The periodic task B (13-1) reads the value written by the periodic task A (11-1) from the buffer 15 and performs an operation using the buffer 15.

The scheduler 14 starts the prediction task 12 at the prediction task start time 24. As a result, the prediction task 12-2 is activated. When the prediction task 12-2 completes the calculation, the prediction task 12-2 writes the prediction value to the buffer 15.

The scheduler 14 activates the periodic task A (11) at the periodic task A start time 25. As a result, the periodic task A (11-2) is started. However, in this example, it is assumed that the calculation of the periodic task A (11-2) is not completed by the predetermined periodic task A deadline 26. The scheduler 14 ends the periodic task A (11-2) at the periodic task A deadline 26.

The scheduler 14 activates the periodic task B (13) at the periodic task B start time 27. As a result, periodic task B (13-2) is activated. The periodic task B (13-2) reads the value written by the prediction task 12-2 from the buffer 15 and performs an operation using the buffer 15.

In other words, the second task unit (processor 2) has the first task (cycled task) by the first task deadline (cycled task A deadline) indicating the deadline timing of the first task (cycled task A) in the execution cycle. If A) is not completed, the second task (periodic task B) is executed by inputting the predicted value of the first task (periodic task A). As a result, even if the first task (cycle task A) is not completed by the first task deadline (cycle task A deadline), the predicted value of the first task (cycle task A) is input as the second task (cycle). Task B) can be completed.

As described above, according to the first embodiment, when a certain process cannot complete the calculation within a predetermined output cycle, the use of the previous value can be prevented by using the output value predicted in advance. It is possible to prevent the processing accuracy from being lowered.

[Example 2]
FIG. 8 is a flowchart showing an example of processing performed by the scheduler 14 according to the second embodiment. The process 500 of the scheduler 14 (processor 2) is executed as follows.
The same description as in FIG. 6 will be omitted.

In step 507, the scheduler 14 compares the time with the predicted task start time β. The scheduler 14 proceeds to step 509 if the time matches the predicted task start time β (step 507: Yes), and if the time does not match the predicted task start time β (step 507: No), step 510. Proceed to.

In step 509, the scheduler 14 determines whether the periodic task B (13) has completed the calculation. If the periodic task B (13) has completed the calculation (step 509: Yes), the process proceeds to step 408. If the periodic task B (13) has not completed the calculation, the process proceeds to step 401.

In step 408, since the time is the predicted task activation time, the scheduler 14 activates the predicted task 12.

In step 510, the scheduler 14 compares the time with the predicted task start time α, and further determines whether the predicted task 12 has not been executed. Here, β <α. If the time matches the predicted task start time α and the predicted task 12 has not been executed (step 510: Yes), the scheduler 14 proceeds to step 511, and the time does not match the predicted task start time α, or the predicted task If 12 has been executed (step 510: No), the process proceeds to step 401.

In step 511, the scheduler 14 activates the prediction task 12.

FIG. 9 is a diagram schematically explaining the operation of each process in the second embodiment on the time axis.

In this example, the case where the prediction task 12-3 is activated at the time (t) = 28 and the periodic task A (11-2) causes an overrun at the time (t) = 26 will be described.
The same description as in FIG. 7 will be omitted.

The scheduler 14 activates the periodic task B (13) at the periodic task B start time 23. As a result, the periodic task B (13-3) is activated.

The scheduler 14 activates the prediction task 12 at the prediction task start time β (28). As a result, the prediction task 12-3 is activated (step 408).

The scheduler 14 does not start the prediction task 12 because the prediction task 12-3 has already been executed at the prediction task start time α (24).

The scheduler 14 activates the periodic task B (13) at the periodic task B start time 27. As a result, the periodic task B (13-4) is activated. The periodic task B (13-4) reads the value written by the prediction task 12-3 from the buffer 15 and performs an operation using the buffer 15.

As described above, according to the second embodiment, the execution time of the processor occupied by the prediction task 12 in the first embodiment can be allocated to any other processing. Further, if the calculation of the periodic task B is completed, the prediction task of the next cycle is started earlier than in the first embodiment.

[Example 3] (The prediction algorithm is changed depending on the amount of free time)
FIG. 10 is a flowchart showing an example of the processing performed in the prediction task 12 according to the third embodiment. The process 600 of the prediction task 12 (processor 2) is executed as follows. The same description as in FIG. 4 will be omitted.

In step 602, it is determined whether or not the prediction task 12 is started at the prediction task start time β.
If the prediction task 12 is started at the prediction task start time β (step 602: Yes), the process proceeds to step 603. If the prediction task 12 is not started at the prediction task start time β (step 602: No), the process proceeds to step 604.

In step 603, the prediction task 12 predicts the value output by the periodic task A (11) based on the acquired input value by using a more accurate prediction algorithm. The highly accurate prediction algorithm is, for example, an algorithm that uses a high-order regression curve or a large amount of data.

In step 604, the prediction task 12 predicts the value output by the periodic task A (11) based on the acquired input value by using a faster prediction algorithm. The high-speed prediction algorithm is, for example, an algorithm that acquires a predicted value corresponding to an input value by referring to a table.

FIG. 11 is a diagram schematically explaining the operation of each process in the third embodiment on the time axis. The same description as in FIGS. 7 and 9 will be omitted.

The scheduler 14 activates the prediction task 12 at the prediction task start time β (28). As a result, the prediction task 12-4 is activated. Since the prediction task 12-4 was started at the prediction task start time β, the prediction is performed using a highly accurate prediction algorithm (step 603).

In other words, the prediction task unit (processor 2) completes the second task (cycle task B) in the execution cycle, and starts predicting the output value of the next first task (cycle task A) (timing). When the free time indicating the time from the predicted task start time β) to the end of the execution cycle exceeds the first threshold value (for example, the time required for the first prediction (high accuracy)), the first priority is given to accuracy over speed. Prediction is performed, and when the free time is equal to or less than the first threshold value, the second prediction is performed in which speed is prioritized over accuracy. As a result, it is possible to switch between the first prediction that prioritizes accuracy and the second prediction that prioritizes speed according to the length of free time.

As described above, according to the third embodiment, the prediction task 12 can output a more accurate prediction value by using a highly accurate prediction algorithm depending on the processing status of other tasks. Further, when the prediction task 12 uses a high-speed prediction algorithm, the calculation load is reduced, so that the power consumed by the vehicle control device can be suppressed.

[Example 4]
FIG. 12 is a flowchart showing an example of the processing performed in the periodic task A (11) according to the fourth embodiment. The process 700 of the periodic task A (processor 2) is executed as follows. The same description as in FIG. 3 will be omitted.

In step 703, the periodic task A determines whether the prediction task skip flag is invalid. If the predictive task skip flag is invalid (step 703: Yes), the process proceeds to step 704. If the predictive task skip flag is not invalid (step 703: No), the process proceeds to step 103.

In step 704, the periodic task A reads the predicted value from the buffer 15.

In step 705, the periodic task A obtains the prediction accuracy using the read predicted value and the calculation result, and compares the prediction accuracy with a predetermined threshold value. If the prediction accuracy is equal to or higher than the threshold value, the process proceeds to step 103. If the prediction accuracy is less than the threshold value, the process proceeds to step 706.

In step 706, the periodic task A enables the predictive task skip flag.

In other words, the first task unit (processor 2) is the prediction task unit (processor 2) when the prediction accuracy indicating the degree of matching between the output value (calculation result) of the periodic task A and the predicted value is lower than the second threshold value. To skip the startup of. As a result, when the prediction accuracy is low, the second task (periodic task B) can be completed by inputting the output value of the first task (periodic task A).

FIG. 13 is a flowchart showing an example of processing performed by the scheduler 14 according to the fourth embodiment. The process 800 of the scheduler 14 (processor 2) is executed as follows. The same description as in FIG. 6 will be omitted.

In step 807, the scheduler 14 determines whether the predictive task skip flag is invalid. If the predictive task skip flag is invalid (step 807: Yes), the process proceeds to step 407. If the predictive task skip flag is not invalid (step 807: No), the process proceeds to step 811.

In step 811 the scheduler 14 subtracts the predicted task skip counter. The predictive task skip counter has a predetermined value.

In step 812, the scheduler 14 determines whether the prediction task skip counter is 0. If it is 0, the process proceeds to step 813. If it is not 0, the process proceeds to step 401.

In step 813, the scheduler 14 invalidates the predictive task skip flag and initializes the value of the predictive task skip counter.

As described above, according to the fourth embodiment, when the accuracy of the prediction does not reach the predetermined threshold value, the execution of the prediction task 12 can be skipped. As a result, the execution time of the processor occupied by the prediction task 12 can be allocated to any other processing. In addition, the electric power consumed by the vehicle control device can be suppressed.

[Example 5]
FIG. 14 is a flowchart showing an example of the processing performed in the periodic task A (11) according to the fifth embodiment. The process 900 of the periodic task A (processor 2) is executed as follows. The same description as in FIG. 3 will be omitted.

In step 902, the periodic task A executes the arithmetic processing I.

In step 903, the periodic task A determines whether the time is within the predetermined arithmetic processing I deadline. If the time is within (previously) the arithmetic processing I deadline (step 903: Yes), the process proceeds to step 904. If the time exceeds the arithmetic processing I deadline (step 903: No), the process proceeds to step 905.

In step 904, the periodic task A executes the arithmetic processing II.

In step 905, the periodic task A generates a predicted task activation event.

In other words, the first task (periodic task A) includes at least the first arithmetic processing (arithmetic processing I) and the second arithmetic processing (arithmetic processing II). When the first arithmetic processing (arithmetic processing I) is completed beyond the first arithmetic processing deadline (arithmetic processing I deadline) indicating the deadline timing, the first task unit (processor 2) performs the second arithmetic processing (arithmetic processing I deadline). The execution of the arithmetic processing II) is canceled, and the prediction task unit (processor 2) is started. As a result, when the first arithmetic processing (arithmetic processing I) is completed beyond the first arithmetic processing deadline (arithmetic processing I deadline), the execution of the second arithmetic processing (arithmetic processing II) is canceled and the second arithmetic processing is canceled. It is not necessary to secure the resources used by the arithmetic processing (arithmetic processing II).

FIG. 15 is a flowchart showing an example of processing performed by the scheduler 14 according to the fifth embodiment. The process 1000 of the scheduler 14 (processor 2) is executed as follows. The same description as in FIG. 6 will be omitted.

In step 1007, the scheduler 14 determines whether or not the predicted task activation event has been received. If the predicted task start event has been received (step 1007: Yes), the process proceeds to step 408. If the predicted task start event has not been received (step 1007: No), the process proceeds to step 401.

FIG. 16 is a diagram schematically explaining the operation of each process in the fifth embodiment on the time axis. The same description as in FIG. 7 will be omitted.

Periodic task A (11-3) completes arithmetic processing I at time 30. However, since this is a time that exceeds the periodic task A arithmetic processing I deadline 29, the predicted task activation event is generated and the processing is terminated (step 905).

Since the scheduler 14 received the prediction task start event at time 30, the scheduler 14 starts the prediction task 12 (step 408). This activates the prediction task 12-5. The prediction task 12-5 predicts the output value of the periodic task A (11-3).

As described above, according to the fifth embodiment, the periodic task A (11) is prevented from overrunning (exceeding the deadline), and the periodic task B13 uses the predicted value output from the predicted task 12. It is possible to perform calculations.

[Example 6]
FIG. 17 is a block diagram showing an example of the configuration of the engine control unit 1-10 to which the present invention according to the sixth embodiment is applied. The same description as in FIG. 1 will be omitted. The sensor input 22 is a device that receives inputs from various sensors such as a crank angle sensor. The ignition driver 23-10 is connected to the ignition system of the engine and flows a current for igniting the plug according to the instruction of the ignition timer 24-10. The ignition timer 24-10 activates the ignition driver 23-10 at a time specified by the processor.

FIG. 18 is a block diagram showing an example of the functional configuration of the engine control unit 1-10 to which the present invention according to the sixth embodiment is applied. The rotation speed calculation task 21-10 counts the crank angle sensor interrupts in the cycle and calculates the rotation speed. The ignition timing calculation task 11-10 calculates the ignition timing from inputs such as the rotation speed and the intake air temperature, and writes the ignition timing value (relative time from the scheduled top dead center arrival time) in the buffer 15. The timer setting task 13-10 reads the ignition timing from the buffer 15 and converts it into an absolute time to set the ignition timer. The prediction task 12 calculates the predicted value of the ignition timing and writes it in the buffer 15. The scheduler 14 starts each task at a predetermined cycle.

FIG. 19 is a flowchart showing an example of processing performed by the ignition timing calculation task 11-10 (processor 2) according to the sixth embodiment. FIG. 20 is a flowchart showing an example of processing performed by the timer set task 13-10 (processor 2) according to the sixth embodiment. The same description as in FIG. 5 will be omitted.

In step 1202 of FIG. 20, the timer setting task sets the ignition timer.

FIG. 21 is a flowchart showing an example of processing performed by the prediction task 12 (processor 2) according to the sixth embodiment. The processing 1300 of the prediction task is executed as follows. The same description as in FIG. 4 will be omitted.

The prediction task 12 inputs the rotation speed in the input process (step 201), and ends the process when the difference between the rotation speed and the previous time is smaller than a certain value (step 1302: Yes). If not (step 1302: Yes), the prediction task 12 determines whether the rotation is high (step 1303), and if it is high rotation (step 1303: Yes), it is a lightweight algorithm, that is, a table reference according to the rotation speed. The referenced value is added to the previous ignition timing to determine the ignition timing.

When the rotation speed is not high (step 1303: No), the prediction task 12 predicts the current rotation speed from a high-precision algorithm, that is, a change in the rotation speed from a plurality of cycles before, and predicts the ignition timing based on the predicted value. (Step 603). The prediction task 12 writes the predicted value of the ignition timing to the buffer 15 in the output process.

In step 1302, the prediction processing and the output processing can be skipped when the difference in the number of rotations from the previous value is smaller than a certain value, and the processing load is increased when the prediction is unnecessary, that is, when the difference from the previous value is less than a certain value. Can be reduced.

In steps 1303, 603, and 604, if the rotation speed is high (step 1303: Yes), the lightweight algorithm (step 604) is performed, and if not (step 1303: No), the prediction process is performed based on the high-precision algorithm (step 1303: No). Step 603), the prediction process can be switched according to the processing load.

As described above, according to the sixth embodiment, when the ignition timing calculation task 11-10 cannot complete the calculation within the predetermined output cycle in the engine control unit, the output value predicted in advance is used to obtain the previous time. The use of the value can be prevented, and the accuracy of the timer set task 13-10 can be prevented from being lowered.

[Example 7]
FIG. 22 is a block diagram showing an example of the configuration of the AD unit 1-20 (Autonomous Driving Unit) to which the present invention is applied. The same description as in FIG. 1 will be omitted.

The memory 3 contains the OS 7, the vertical position estimation task 11-20, the horizontal position estimation task 11-21, the vertical position prediction task 12-20, the horizontal position prediction task 12-21, and the scene determination task 31. Fusion task 32 is loaded and executed by processor 2.

The AD unit is a control unit that controls automatic driving, receives inputs from various sensors such as cameras, recognizes objects around the vehicle from the inputs, and integrates (fusion) the recognized object information. , After determining the driving situation (scene) of the own vehicle and estimating the self-position of the own vehicle, calculate the route that the own vehicle should take and control the steering angle etc. required to travel on that route. The information is calculated and the control information is output to an actuator such as an EPS (electric steering system).

Since the present invention applies to self-position estimation among the functions of the AD unit, only the components related to self-position estimation are shown.

FIG. 23 is a block diagram showing an example of the functional configuration of the AD unit 1-20 to which the present invention is applied.

The fusion task 32 integrates the recognized object information and the like, and outputs the information required by the vertical position estimation task 11-20 and the horizontal position estimation task 11-21.

The scene determination task 31 determines whether the risk of collision with the preceding vehicle is above a certain level while the own vehicle is accelerating or decelerating the preceding vehicle, or whether the left and right lanes are being changed and output.

The vertical position estimation task 11-20 estimates the vertical position of the own vehicle and outputs the result to the vertical position buffer 15-20.

The horizontal position estimation task 11-21 estimates the horizontal position of the own vehicle and outputs the result to the horizontal position buffer 15-21.

The vertical position prediction task 12-20 predicts the vertical position of the own vehicle and outputs the result to the vertical position buffer 15-20.

The horizontal position prediction task 12-21 predicts the horizontal position of the own vehicle and outputs the result to the horizontal position buffer 15-21.

Route calculation task 13-20 calculates the route that the vehicle should take from the vertical and horizontal positions of the vehicle, lane information, free space information, and the like.

FIG. 24 is a flowchart showing an example of the processing performed in the vertical position estimation task 11-20 according to the seventh embodiment. FIG. 25 is a flowchart showing an example of the processing performed in the horizontal position estimation task 11-21 according to the seventh embodiment. FIG. 26 is a flowchart showing an example of the processing performed in the route calculation task 13-20 according to the seventh embodiment. The process of the route calculation task 1600 is executed as follows. The same description as in FIG. 5 will be omitted.

In step 1601 of FIG. 26, the input value is acquired from the vertical position buffer 15-20 and the horizontal position buffer 15-21.

FIG. 27 is a flowchart showing an example of the processing performed in the vertical position prediction task 12-20 according to the seventh embodiment. The process 1700 of the vertical position prediction task (processor 2) is executed as follows. The same description as in FIG. 4 will be omitted.

In step 1702, the vertical position prediction task determines whether there is a risk of the own vehicle colliding with the preceding vehicle.

The vertical position prediction task 12-20 inputs the fusion result and the scene determination result in the input process (step 201), and ends the process when the collision risk with the preceding vehicle is below a certain level (step 1702: No). If not (step 1702: Yes), the vertical position of the own vehicle is predicted in the prediction process (step 202), and the output value is written to the vertical position buffer 15-20 in the output process (step 203).

FIG. 28 is a flowchart showing an example of the processing performed in the horizontal position prediction task 12-21 according to the seventh embodiment. The process 1800 of the horizontal position prediction task (processor 2) is executed as follows. The same description as in FIG. 4 will be omitted.

In step 1802, the lateral position prediction task 12-21 determines whether the lane is being changed.

The horizontal position prediction task 12-21 inputs the fusion result and the scene determination result in the input process (step 201), and ends the process when the lane is not being changed (step 1802: No). If not (step 1802: Yes), the vehicle lateral position is predicted in the prediction process (step 202), and the output value is written to the lateral position buffer 15-21 in the output process (step 203).

In steps 1702 and 1802, it is possible to switch which prediction processing is to be performed based on the scene determination result, and it is possible to avoid an increase in processing load due to unnecessary prediction processing.

As described above, according to the seventh embodiment, when the vertical position estimation task 11-20 or the horizontal position estimation task 11-21 cannot complete the calculation within the predetermined output cycle in the AD unit, the output predicted in advance. By using the value, it is possible to prevent the use of the previous value and prevent the accuracy of the route calculation task 13-20 from being lowered.

The present invention is not limited to the above-described examples, and includes various modifications.
For example, the above-described examples have been described in detail in order to explain the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the described configurations. Further, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Further, it is possible to add / delete / replace a part of the configuration of each embodiment with another configuration.

Further, each of the above configurations, functions, etc. may be realized by hardware, for example, by designing a part or all of them with an integrated circuit. Further, each of the above configurations, functions, and the like may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function can be placed in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

The embodiment of the present invention may have the following aspects.

(1). In a control device including a task execution unit (processor) that executes a task in a predetermined cycle, the task execution unit is characterized in that the output value of the task in the execution cycle is predicted before the output in the execution cycle. Control device.

(2). In (1), the task execution unit is a control device that makes it possible to receive the predicted value when the task is not completed within the execution cycle.

(3). In (1) or (2), when the task execution unit predicts the output value in the execution cycle, if the free time of any of the execution cycles before the output in the execution cycle is longer than a predetermined value, the accuracy is correct. A control device that makes a prediction that gives priority to speed, and makes a prediction that gives priority to speed when the free time of any of the execution cycles before output in the execution cycle is less than a predetermined value.

(4). In (1) or (2), when the delay of the task in the execution cycle exceeds the threshold value, the task execution unit interrupts the task in the execution cycle and predicts the output value in the execution cycle. A control device characterized by that.

(5). In (3), the prediction that prioritizes accuracy uses an algorithm that uses a high-order regression curve or a large amount of data, and the prediction that prioritizes speed uses an algorithm that acquires the predicted value corresponding to the input value by referring to the table. Use.

(6). In (1), the control device is an AD unit.

That is, the control device is a control device that has a processor and a memory, performs calculations using input data acquired from the input device, generates an output value, and controls the output device within a predetermined time. A periodic task A that receives data from an input device and performs predetermined arithmetic processing on the received data to generate an output value, and a periodic task B that performs arithmetic processing using the generated output value as an input. It has a prediction task that predicts the output value of the cycle task A, and the cycle task B uses the output value of the prediction task when the cycle task A does not complete the process within a predetermined time. , Perform arithmetic processing.

According to this control device, when a certain process cannot complete the calculation within a predetermined output cycle, it is possible to prevent the processing accuracy from being lowered by using the output value predicted in advance.

1 ... Vehicle control device 2 ... Processor 3 ... Memory 4 ... Non-volatile memory 8 ... Interconnect 11 ... Periodic task A
12 ... Prediction task 13 ... Periodic task B
14 ... Scheduler 15 ... Buffer

Claims (6)

1. With a buffer
   A first task unit that executes the first task in a predetermined execution cycle and writes the output value of the first task to the buffer.
   A prediction task unit that predicts the output value of the first task and writes the predicted value to the buffer before the output value of the first task is written to the buffer in the execution cycle.
   In the execution cycle, the output value or the predicted value of the first task is read from the buffer, the second task is executed by inputting the output value or the predicted value read from the buffer, and the output of the second task is output. The second task part that outputs the value and
   A control device characterized by being equipped with.
2. The control device according to claim 1.
   The second task part is
   When the first task is not completed by the first task deadline indicating the deadline timing of the first task in the execution cycle, the second task is executed by inputting the predicted value of the first task. Control device.
3. The control device according to claim 2.
   The prediction task section
   The free time indicating the time from the timing at which the second task is completed in the execution cycle and the timing at which the prediction of the output value of the next first task is started to the timing at which the execution cycle ends exceeds the first threshold value. In the case, make the first prediction that prioritizes accuracy over speed,
   A control device characterized in that when the free time is equal to or less than the first threshold value, a second prediction is performed in which speed is prioritized over accuracy.
4. The control device according to claim 2.
   The first task is
   Including at least the first arithmetic processing and the second arithmetic processing,
   The first task section is
   A control device characterized in that when the first arithmetic processing is completed beyond the deadline of the first arithmetic processing indicating the deadline timing, the execution of the second arithmetic processing is canceled and the prediction task unit is activated.
5. The control device according to claim 2.
   The first task section is
   A control device characterized in that when the prediction accuracy indicating the degree of coincidence between the output value and the predicted value is lower than the second threshold value, the activation of the predicted task unit is skipped.
6. The first step of executing the first task in a predetermined execution cycle and writing the output value of the first task to the buffer, and
   A second step of predicting the output value of the first task and writing the predicted value to the buffer before the output value of the first task is written to the buffer in the execution cycle.
   In the execution cycle, the output value or the predicted value of the first task is read from the buffer, the second task is executed by inputting the output value or the predicted value read from the buffer, and the output of the second task is output. The third step to output the value and
   How to get the processor to run.

Images