Health monitoring method, system and vehicle for autonomous driving controller
Through configuration file processing, status data collection and abnormal state processing methods, the key parameters of the autonomous driving controller are monitored in real time, solving the problem of lack of health monitoring in the existing technology, and achieving the safety and stability of autonomous driving cars. Monitor and manage.
Patent Information
- Application Number
- CN202110810400.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-07-19
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-07-19
AI Technical Summary
The prior art lacks effective health monitoring methods for autonomous driving controllers, and cannot detect and remind users of abnormal controllers in real time, which may lead to car safety hazards.
Through the steps of configuration file processing, status data acquisition, logic analysis and abnormal state processing, the CPU temperature, CPU load, memory load and memory life cycle are monitored in real time, the abnormal state is judged, and the fault code is reported through the CAN bus or on-board Ethernet ( DTC) so that users can turn off autonomous driving in a timely manner.
Real-time detection and judgment of system-level and process-level abnormal states is realized, the health management level of autonomous driving cars is improved, and driving safety and system stability are ensured.
Smart Images

Figure CN115701891B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of automobile monitoring technology, and specifically relates to a health monitoring method, system and vehicle for an autonomous driving controller. Background Art
[0002] With the development of control technology and information technology, more and more automobile manufacturers have begun to use autonomous driving technology. The autonomous driving controller is the brain of the autonomous driving car, involving route planning and decision-making during the driving process, comprehensive management of cars and roads, and many other aspects.
[0003] The autopilot controller is the core of the self-driving car. The health status of the autopilot controller during operation is crucial. Currently, there is no health monitoring method for the autopilot controller to maintain the health status of the autopilot controller and remind the user that the autopilot controller is in an abnormal state and needs to shut down the autopilot.
[0004] Therefore, it is necessary to develop a health monitoring method, system and vehicle for an autonomous driving controller. Summary of the invention
[0005] The purpose of the present invention is to provide a health monitoring method, system and vehicle for an autonomous driving controller.
[0006] A health monitoring method for an autonomous driving controller according to the present invention comprises the following steps:
[0007] Step 1. Configuration file processing: Get the parameters of each monitoring sub-item from the configuration file;
[0008] Step 2. Status data collection: real-time collection of data on CPU temperature, total CPU load, total memory load, memory life cycle, as well as real-time collection of data on single process memory load and single process CPU load;
[0009] Step 3. Logical analysis: Based on the data from step 2, analyze and calculate whether the monitored items are in an abnormal state. Abnormal states include high CPU temperature, high total CPU load, high total memory load, high CPU load of a single process, high memory load of a single process, memory leak of a single process, and memory life cycle exhaustion.
[0010] Step 4. Abnormal status processing: When the monitoring item is abnormal, report it to DTC.
[0011] Optionally, in the first step, the parameters include an original data sampling period, a data unit, an abnormal state triggering threshold, and a normal state recovery threshold.
[0012] Optionally, in the second step, data on CPU temperature, total CPU load, and total memory load are obtained through a standard system call interface; data on the CPU load of a single process is obtained by calculating the CPU time occupied by a single process; data on the memory load of a single process is obtained by calculating the address space of the stack segment, code segment, and data segment occupied by a single process; data on the memory life cycle is obtained by reading the register value of the memory.
[0013] Optionally, in the third step, the sampling period of the CPU temperature is once every 2 seconds, the data unit is "℃", the abnormal state trigger threshold is more than 110℃ for 2 consecutive times, and the normal state recovery threshold is less than 100℃ for 2 consecutive times; the sampling period of the total CPU load is once every 2 seconds, the data unit is "%", the abnormal state trigger threshold is more than 90% for 30 consecutive times, and the normal state recovery threshold is less than 90% for 30 consecutive times; the sampling period of the total memory load is once every 2 seconds, the data unit is "%", the abnormal state trigger threshold is more than 90% for 30 consecutive times, and the normal state recovery threshold is less than 90% for 30 consecutive times; the sampling period of the memory life cycle is 180 The sampling period of the CPU load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is exceeding 90% once. The memory life cycle is irreversible and there is no recovery threshold. The sampling period of the CPU load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is exceeding 50% for 12 consecutive times, and the normal state recovery threshold is less than 50% for 12 consecutive times. The sampling period of the memory load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is exceeding 50% for 12 consecutive times, and the normal state recovery threshold is less than 50% for 12 consecutive times. The memory load of a single process has been increasing for 60 consecutive times, triggering a memory leak.
[0014] Optionally, in the fourth step, DTC is reported via the CAN bus or the vehicle Ethernet.
[0015] In a second aspect, a health monitoring system for an autonomous driving controller according to the present invention includes a memory and a controller;
[0016] The memory stores a computer-readable program;
[0017] When the controller calls the computer-readable program, it can execute the steps of the health monitoring method of the autonomous driving controller as described in the present invention.
[0018] In a third aspect, a vehicle described in the present invention adopts a health monitoring system of the automatic driving controller described in the present invention.
[0019] The present invention has the following advantages: Based on system status information monitoring, the present invention can detect and judge abnormal conditions at the system level and process level in real time with very low computational cost, solves the health monitoring problem of the autonomous driving controller, and has high application value in health management in the professional field of autonomous driving vehicle monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 is a principle block diagram of the automatic driving controller in this embodiment;
[0021] Figure 2 is a flow chart of this embodiment. DETAILED DESCRIPTION
[0022] The present invention is described in detail below with reference to the accompanying drawings.
[0023] like Figure 1 As shown in the figure, the core of the automatic driving controller includes CPU1, memory 2 and storage 3, and the peripheral interfaces are CAN bus 4 and vehicle Ethernet 5. The function of CPU1 is to process instructions and execute operations, and is responsible for the route planning and decision-making of automatic driving; the function of memory 2 is the temporary data storage medium of the running program; the function of storage 3 is data storage, which has a limit on the number of reads and writes; CAN bus 4 is the standard bus for internal communication in the car, and is the outlet for reporting DTC; vehicle Ethernet 5 is an emerging communication technology inside the car, with a faster transmission speed, and is also the outlet for reporting DTC.
[0024] like Figure 2 As shown, in this embodiment, a health monitoring method for an autonomous driving controller includes the following steps:
[0025] Step 1. Configuration file processing: Get the parameters of each monitoring sub-item from the configuration file.
[0026] Step 2. Status data collection: real-time collection of data on CPU temperature, total CPU load, total memory load, and memory life cycle, as well as real-time collection of data on single process memory load and single process CPU load.
[0027] Step 3. Logical analysis: Based on the data from the second step, analyze and calculate whether the monitored items are in an abnormal state. Abnormal states include CPU temperature too high, total CPU load too high, total memory load too high, single process CPU load too high, single process memory load too high, single process memory leak, and memory life cycle exhaustion.
[0028] Step 4. Abnormal status processing: When the monitored item is abnormal, DTC is reported to the user via the CAN bus or the vehicle Ethernet. In this embodiment, different DTCs are reported via the CAN bus or the vehicle Ethernet according to different error types.
[0029] In this embodiment, in the first step, the parameters include the original data sampling period, the data unit, the abnormal state triggering threshold and the normal state recovery threshold.
[0030] In this embodiment, in the second step, the data of CPU temperature, total CPU load, and total memory load are obtained through the standard system call interface; the data of CPU load of a single process is obtained by calculating the CPU time occupied by a single process; the data of memory load of a single process is obtained by calculating the address space of the stack segment, code segment, and data segment occupied by a single process; the data of the memory life cycle is obtained by reading the register value of the memory.
[0031] In this embodiment, in the third step, the sampling period of CPU temperature is once every 2 seconds, the data unit is "℃", the abnormal state trigger threshold is twice consecutively exceeding 110℃, and the normal state recovery threshold is twice consecutively below 100℃; the sampling period of CPU total load is once every 2 seconds, the data unit is "%", the abnormal state trigger threshold is 30 consecutive times exceeding 90% (full load 100%), and the normal state recovery threshold is 30 consecutive times below 90%; the sampling period of total memory load is once every 2 seconds, the data unit is "%", the abnormal state trigger threshold is 30 consecutive times exceeding 90% (full load 100%), and the normal state recovery threshold is 30 consecutive times below 90%; the sampling period of memory life cycle is 180 The sampling period of the CPU load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is exceeding 90% once. The memory life cycle is irreversible and there is no recovery threshold. The sampling period of the CPU load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is exceeding 50% (full load 100%) for 12 consecutive times, and the normal state recovery threshold is less than 50% for 12 consecutive times. The sampling period of the memory load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is exceeding 50% (full load 100%) for 12 consecutive times, and the normal state recovery threshold is less than 50% for 12 consecutive times. The memory load of a single process has been increasing for 60 consecutive times, triggering a memory leak.
[0032] In the present embodiment, a health monitoring system for an autonomous driving controller includes a memory and a controller; the memory stores a computer-readable program; when the controller calls the computer-readable program, it can execute the steps of the health monitoring method for the autonomous driving controller as described in the present embodiment.
[0033] In this embodiment, a vehicle adopts the health monitoring system of the autonomous driving controller as described in this embodiment.
Claims
1. A health monitoring method for an autonomous driving controller, It is characterized in that The following steps are involved: Step 1. Configuration file processing: Get the parameters of each monitoring sub-item from the configuration file; the parameters include the original data sampling period, data unit, abnormal state trigger threshold and normal state recovery threshold; Step 2. Status data collection: real-time collection of data on CPU temperature, total CPU load, total memory load, memory life cycle, as well as real-time collection of data on single process memory load and single process CPU load; Step 3. Logical analysis: Based on the data from the second step, analyze and calculate whether the monitoring item is in an abnormal state. Abnormal states include high CPU temperature, high total CPU load, high total memory load, high CPU load of a single process, high memory load of a single process, memory leak of a single process, and memory life cycle exhaustion; the sampling period of CPU temperature is once every 2 seconds, and the data unit is "℃". The abnormal state trigger threshold is 2 consecutive times exceeding 110℃, and the normal state recovery threshold is 2 consecutive times below 100℃; the sampling period of total CPU load is once every 2 seconds, and the data unit is "%". The abnormal state trigger threshold is 30 consecutive times exceeding 90%, and the normal state recovery threshold is 30 consecutive times below 90%; the sampling period of total memory load is once every 2 seconds, and the data unit is "%". The state trigger threshold is 30 times in a row exceeding 90%, and the normal state recovery threshold is 30 times in a row below 90%; the sampling period of the memory life cycle is once every 1800 seconds, and the data unit is "%". The abnormal state trigger threshold is once exceeding 90%. The memory life cycle is irreversible and there is no recovery threshold; the sampling period of the CPU load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is 12 times in a row exceeding 50%, and the normal state recovery threshold is 12 times in a row below 50%; the sampling period of the memory load of a single process is once every 5 seconds, and the data unit is "%". The abnormal state trigger threshold is 12 times in a row exceeding 50%, and the normal state recovery threshold is 12 times in a row below 50%. The memory load of a single process has been increasing for 60 times in a row, triggering a memory leak; Step 4. Abnormal status processing: When the monitoring item is abnormal, report it to DTC.
2. The health monitoring method of the autonomous driving controller according to claim 1, Features: In the second step, the data of CPU temperature, total CPU load, and total memory load are obtained through the standard system call interface; the data of CPU load of a single process is obtained by calculating the CPU time occupied by a single process; The data of memory load of a single process is obtained by calculating the address space of the stack segment, code segment, and data segment occupied by a single process; the data of memory life cycle is obtained by reading the register value of the memory.
3. The health monitoring method of the automatic driving controller according to claim 1 or 2, Features: In the fourth step, DTC is reported via the CAN bus or the vehicle Ethernet.
4. A health monitoring system for an autonomous driving controller, Features: including a memory and a controller; The memory stores a computer-readable program; When the controller calls the computer-readable program, it can execute the steps of the health monitoring method of the autonomous driving controller as described in any one of claims 1 to 3.
5. A vehicle, Features: A health monitoring system for an autonomous driving controller as claimed in claim 4 is adopted.
