Safe brain controller communication execution architecture for automatic driving

Abstract

The invention discloses a safe brain controller communication execution architecture for automatic driving, and particularly relates to the field of automatic driving, which comprises a data acquisition unit, a sensing unit, a planning unit, a control unit, a safety supervision unit and a control execution unit, the data acquisition unit comprises vehicle sensing data; networking the V2X data; vehicle body data; power data; map data; when the data acquisition unit transmits the data to the intelligent driving algorithm unit, the data acquisition unit transmits the data to the vehicle-mounted safety supervision unit at the same time for online safety verification. According to the invention, for a sensing result, the accident experience module is used for matching a sensing scene in the monitoring module, and when a similar scene exists, the intelligent driving system is reminded of a safety risk.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and more specifically, to a safe brain controller communication execution architecture for autonomous driving. Background Art

[0002] With the development of smart cars, advanced autonomous driving has become an indispensable function for current vehicles due to its convenience and safety. People are increasingly relying on assisted driving or autonomous driving functions to reduce the burden of driving and improve driving comfort and safety.

[0003] In the current in-vehicle intelligent computing platform, the intelligent driving algorithm is integrated into the corresponding domain controller, and only basic data transmission integrity tests are performed during the internal operation of the algorithm. For example, for the image sensor in the perception system, the algorithm only detects the integrity of the sensor output data, and when the output data is complete, the perception result is transmitted to the downstream sensor. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a safety brain controller communication execution architecture for autonomous driving. Through an on-board safety supervision brain, the sensor data, decision-making planning data, and trajectory control data during the operation of the autonomous driving vehicle are supervised, forming a highly reliable driving data supervision method, providing safety warnings for dangerous driving scenarios, and improving the safety of intelligent vehicle operation.

[0005] To achieve the above-mentioned object, the present invention provides the following technical solutions: a communication execution architecture of a safety brain controller for autonomous driving, comprising: a data acquisition unit, a perception unit, a planning unit, a control unit, a safety supervision unit and a control execution unit; The data collection unit includes: vehicle perception data; networked V2X data; body data; power data; map data; When the data from the data collection unit is transmitted to the intelligent driving algorithm unit, it will also transmit the data to the on-board safety supervision unit for online safety verification; The data in the perception unit, planning unit and control unit will be transmitted to the safety supervision unit, and the data in the safety supervision unit will be transmitted to the corresponding unit for algorithm supervision and accident prevention; The safety supervision unit is a safety supervision unit driven by a knowledge engine, which includes an accident experience supervision module, a common sense supervision module and a data recording module.

[0006] Furthermore, the accident scene data set is preset in the accident experience landscape module, and after the perception sensor data is input into the accident experience supervision module, scene similarity calculation is performed therein for the prevention of accident scenes and the escape from passive dangerous scenes.

[0007] Furthermore, the common sense supervision module has scene reasoning supervision function, algorithm state supervision function, security soft rule supervision function, and security hard rule supervision function.

[0008] Furthermore, the scenario reasoning function is used to monitor the state of the intelligent vehicle operation design domain.

[0009] Furthermore, the algorithm status supervision function is used to supervise the algorithm status in the perception unit, planning unit, and control unit, and to supervise the traffic light status information in the v2x signal.

[0010] Furthermore, the safety soft rule supervision function is used to supervise the comfort data during vehicle operation.

[0011] Furthermore, the safety hard rule function is used to monitor whether smart vehicles have any illegal crossing of the line or running of red lights.

[0012] Furthermore, all the supervision data will be stored in the data recording module.

[0013] A safety brain controller communication execution architecture for autonomous driving includes the following steps: Step 1: After the smart driving vehicle is started, assisted driving or automatic driving is turned on, and the data acquisition unit records and stores data such as cameras, radars, V2X, IMU, GPS, GNSS, vehicle body, power, and maps; Step 2: Check the data integrity in the safety supervision unit. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 3: In the perception unit, the perception data is identified, including the position and speed of objects such as vehicles, pedestrians, and obstacles in the environment; Step 4: Based on the scene perception results, the safety supervision unit matches the dangerous accident scene. When a similar scene appears, the safety supervision unit reminds the vehicle that there is a danger; Step 5: Check the data integrity of the scene sensing results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 6: The perception results are used in the decision-making and planning unit to make driving decisions and plan paths; Step 7: Check the data integrity of the decision-making planning results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 8: Based on the decision-making planning results, the safety supervision unit makes a safe driving experience judgment. When there are abnormal driving decisions, the safety supervision unit provides vehicle guidance suggestions; Step 9: Check the data integrity of the decision-making planning results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 10: Calculate the intelligent driving vehicle control instructions in the trajectory control unit.

[0014] Step 11: Check the data integrity of the trajectory control result. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 12: Check the control command for safety, comfort, and compliance. When the check is normal, execute the algorithm control command. When the check is abnormal, forcibly take over the vehicle.

[0015] Technical effects and advantages of the present invention: Through the knowledge engine supervision module, the driving status of the intelligent vehicle is monitored based on the accident experience unit and the common sense supervision unit, including the monitoring of the algorithm status, the monitoring of the operation design domain and the monitoring of rationality, so as to establish a redundant intelligent driving safety protection system, achieve the effect of real-time recognition and prevention of driving risks, and enhance users' trust in the intelligent driving system.

[0016] At the same time, when a driving accident occurs, the data in the knowledge engine supervision module can be used as third-party data to provide factual basis for accident supervision and tracing. At the same time, the data in the accident experience unit can be used as reference data to provide training data for intelligent driving algorithms to improve driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is the overall data flow diagram of the present invention.

[0018] Figure 2 This is a unit module diagram of the present invention.

[0019] Figure 3 This is a block diagram of the security supervision unit module driven by the knowledge engine of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] according to Figure 1-3 A safety brain controller communication execution architecture for autonomous driving is shown, comprising the following steps: Step 1: After the smart driving vehicle is started, assisted driving or automatic driving is turned on, and the data acquisition unit records and stores data such as cameras, radars, V2X, IMU, GPS, GNSS, vehicle body, power, and maps; Step 2: Check the data integrity in the safety supervision unit. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 3: In the perception unit, the perception data is identified, including the position and speed of objects such as vehicles, pedestrians, and obstacles in the environment; Step 4: Based on the scene perception results, the safety supervision unit matches the dangerous accident scene. When a similar scene appears, the safety supervision unit reminds the vehicle that there is a danger; Step 5: Check the data integrity of the scene sensing results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 6: The perception results are used in the decision-making and planning unit to make driving decisions and plan paths; Step 7: Check the data integrity of the decision-making planning results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 8: Based on the decision-making planning results, the safety supervision unit makes a safe driving experience judgment. When there are abnormal driving decisions, the safety supervision unit provides vehicle guidance suggestions; Step 9: Check the data integrity of the decision-making planning results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 10: Calculate the intelligent driving vehicle control instructions in the trajectory control unit.

[0022] Step 11: Check the data integrity of the trajectory control result. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 12: Check the control command for safety, comfort, and compliance. When the check is normal, execute the algorithm control command. When the check is abnormal, forcibly take over the vehicle.

[0023] When the data from the data acquisition unit is transmitted to the intelligent driving algorithm unit, the data will also be transmitted to the on-board safety supervision unit for online safety verification. The data in the perception unit, planning unit and control unit will be transmitted to the safety supervision unit, and the data in the safety supervision unit will be transmitted to the corresponding unit for algorithm supervision and accident prevention. The safety supervision unit is a safety supervision unit driven by a knowledge engine, which includes an accident experience supervision module, a common sense supervision module and a data recording module. The accident scene data set is preset in the accident experience landscape module. After the perception sensor data is input into the accident experience supervision module, the scene similarity calculation is performed in it for accident scene prevention and passive danger scene escape. The common sense supervision module has scene reasoning supervision function, algorithm state supervision function, safety soft rule supervision function and safety hard rule supervision function. The scene reasoning function is used to supervise the state of the intelligent vehicle operation design domain, and the algorithm state supervision function is used to supervise the algorithm state in the perception unit, planning unit and control unit, and supervise the traffic light state in the v2x signal and other information; the safety soft rule function is used to supervise the comfort data in vehicle operation, such as emergency braking and emergency steering data. The safety hard rule function is used to monitor whether smart vehicles violate the rules, such as crossing the line or running a red light. All monitoring data will be stored in the data recording module.

Claims

1. A safe brain controller communication execution architecture for autonomous driving, characterized in that: include: Data collection unit, perception unit, planning unit, control unit, safety supervision unit and control execution unit; The data collection unit includes: vehicle perception data; networked V2X data; body data; power data; map data; When the data from the data collection unit is transmitted to the intelligent driving algorithm unit, it will also transmit the data to the on-board safety supervision unit for online safety verification; The data in the perception unit, planning unit and control unit will be transmitted to the safety supervision unit, and the data in the safety supervision unit will be transmitted to the corresponding unit for algorithm supervision and accident prevention; The safety supervision unit is a safety supervision unit driven by a knowledge engine, which includes an accident experience supervision module, a common sense supervision module and a data recording module.

2. A safe brain controller communication execution architecture for autonomous driving according to claim 1, characterized in that: The accident experience landscape module is pre-set with an accident scene data set. After the perception sensor data is input into the accident experience supervision module, scene similarity calculation is performed therein for preventing accident scenes and getting rid of passive dangerous scenes.

3. The communication execution architecture of a safety brain controller for autonomous driving according to claim 1, characterized in that: The common sense supervision module has scene reasoning supervision function, algorithm state supervision function, security soft rule supervision function, and security hard rule supervision function.

4. The communication execution architecture of a safety brain controller for autonomous driving according to claim 3, characterized in that: The scenario reasoning function is used to monitor the state of the intelligent vehicle operation design domain.

5. The communication execution architecture of a safe brain controller for autonomous driving according to claim 3, characterized in that: The algorithm status supervision function is used to supervise the algorithm status in the perception unit, planning unit, and control unit, and at the same time supervise the traffic light status information in the v2x signal.

6. The communication execution architecture of a safe brain controller for autonomous driving according to claim 3, characterized in that: The safety soft rule supervision function is used to supervise the comfort data during vehicle operation.

7. The communication execution architecture of a safety brain controller for autonomous driving according to claim 3, characterized in that: The safety hard rule function is used to monitor whether smart vehicles violate the rules by crossing the line or running a red light.

8. The communication execution architecture of a safe brain controller for autonomous driving according to claim 1, characterized in that: All the supervision data will be stored in the data recording module.

9. The communication execution architecture of a safe brain controller for autonomous driving according to claim 1, characterized in that: The following steps are involved: Step 1: After the smart driving vehicle is started, assisted driving or automatic driving is turned on, and the data acquisition unit records and stores data such as cameras, radars, V2X, IMU, GPS, GNSS, vehicle body, power, and maps; Step 2: Check the data integrity in the safety supervision unit. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 3: In the perception unit, the perception data is identified, including the position and speed of objects such as vehicles, pedestrians, and obstacles in the environment; Step 4: Based on the scene perception results, the safety supervision unit matches the dangerous accident scene. When a similar scene appears, the safety supervision unit reminds the vehicle that there is a danger; Step 5: Check the data integrity of the scene sensing result. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 6: The perception results are used in the decision-making and planning unit to make driving decisions and plan paths; Step 7: Check the data integrity of the decision-making planning results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 8: Based on the decision-making planning results, the safety supervision unit makes a safe driving experience judgment. When there are abnormal driving decisions, the safety supervision unit provides vehicle guidance suggestions; Step 9: Check the data integrity of the decision-making planning results. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 10: Calculating intelligent driving vehicle control instructions in the trajectory control unit; Step 11: Check the data integrity of the trajectory control result. If the data integrity is abnormal, instruct the vehicle to stop compulsorily. Step 12: Check the control command for safety, comfort, and compliance. When the check is normal, execute the algorithm control command. When the check is abnormal, forcibly take over the vehicle.

Images