Risk Mitigation Control Method, System and Vehicle for Human-Machine Cooperative Driving

Abstract

The present invention discloses a risk mitigation control method, system and vehicle for human-machine co-driving, comprising the following steps: Step 1. Build a human-machine co-driving model oriented to safety; Step 2. Real-time monitor the driver takeover ability and the control ability of the automatic driving system and input them into the human-machine co-driving model. The human-machine co-driving model calculates the safety state slipping path according to the input signals, and outputs the risk mitigation measure trigger requests and the priorities of the responses for each safety state slipping path; Step 3. Collect all the risk mitigation measure trigger requests under the safety state slipping paths, and respond to the risk mitigation measure trigger request with the highest priority. The present invention can effectively evaluate the safety of human-machine co-driving, and comprehensively give different risk mitigation strategies according to the safety of humans and machines.

Description

Technical Field

[0001] The present invention belongs to the technical field of autonomous driving, and particularly relates to a risk mitigation control method, system and storage medium for human-machine co-driving. Background Art

[0002] The risk mitigation strategies for autonomous driving mainly adopt measures such as takeover alarm, safe parking, function degradation, and active collision avoidance. When the autonomous driving system anticipates that it cannot drive safely, it requests the driver to take over the vehicle control right and perform manual driving. Currently, there are many patents on system control and driver monitoring in human-machine co-driving, such as CN201910700881.0 Driving authority switching system considering driver state in human-machine co-driving environment, CN201910721212.1 Abnormal state monitoring system for driver during driving authority switching in human-machine co-driving environment, CN201911358599.5 Dynamic human-machine co-driving driving right allocation method based on driver's real-time risk response, and so on. These methods propose specific system architectures and software algorithms for the handover method, handover timing, and handover process of vehicle control rights, but cannot effectively evaluate the safety of human-machine co-driving. Summary of the Invention

[0003] The purpose of the present invention is to provide a risk mitigation control method, system and storage medium for human-machine co-driving, which can effectively evaluate the safety of human-machine co-driving and comprehensively give different risk mitigation strategies according to the safety of humans and machines.

[0004] In a first aspect, a risk mitigation control method for human-machine co-driving according to the present invention includes the following steps:

[0005] Step 1. Build a safety-oriented human-machine co-driving model:

[0006] Take the weakening degree and weakening speed of the driver's takeover ability, and the weakening degree and weakening speed of the autonomous driving system control ability as input signals;

[0007] Take the safety measures corresponding to the safety state sliding path as output signals;

[0008] Establish a human-machine co-driving model between the input signals and the output signals; wherein, the human-machine co-driving model is configured to: calculate the safety state sliding path according to the weakening degree and weakening speed of the driver's takeover ability, and the weakening degree and weakening speed of the autonomous driving system control ability, and output the risk mitigation measure trigger request and response priority corresponding to the safety state sliding path;

[0009] Step 2. Monitor the driver takeover ability and the automatic driving system control ability in real time and input them into the human-machine co-driving model. The human-machine co-driving model calculates the safety state slipping path based on the input signals, and outputs the risk mitigation measure trigger requests and the response priorities for each safety state slipping path;

[0010] Step 3. Collect all the risk mitigation measure trigger requests under the safety state slipping paths, and respond to the risk mitigation measure trigger request with the highest priority.

[0011] Optionally, the driver takeover ability is distributed between "having takeover ability" and "not having takeover ability"; the ability of the automatic driving system to control the vehicle is distributed between "having control ability" and "not having control ability";

[0012] Define the situation where the driver has takeover ability and the automatic driving system has control ability as a safe state;

[0013] Define the situation where the driver does not have takeover ability and the automatic driving system has control ability as the first safe state;

[0014] Define the situation where the driver has takeover ability and the automatic driving system does not have control ability as the second safe state;

[0015] Define the situation where the driver does not have takeover ability and the automatic driving system does not have control ability as an unsafe state;

[0016] Define the transition from the safe state to the first safe state as the first safe state slipping path;

[0017] Define the transition from the safe state to the second safe state as the second safe state slipping path;

[0018] Define the transition from the first safe state to the unsafe state as the third safe state slipping path;

[0019] Define the transition from the second safe state to the unsafe state as the fourth safe state slipping path;

[0020] Define the transition from the safe state to the unsafe defined state as the fifth safe state slipping path.

[0021] Optionally, the risk mitigation measures corresponding to the first safe state slipping path and the second safe state slipping path are to arouse the driver's attention through takeover alarms and inform the driver to immediately take over the vehicle;

[0022] The risk mitigation measures corresponding to the third safe state slipping path are to reduce the vehicle speed by means of comfortable braking and continuously issue a stronger takeover alarm to remind the driver to take over as soon as possible; if there is an emergency collision risk or the dangerous situation exceeds the scope that the driver can handle, then adopt the emergency braking method, and the autonomous driving system can control the vehicle to steer;

[0023] The risk mitigation measures corresponding to the fourth safe state slipping path are that the driver is required to complete the takeover immediately. During this process, the autonomous driving system needs to continue to control the vehicle movement using the remaining control capabilities. If the autonomous driving system still has the braking ability, then adopt comfortable braking or emergency braking according to the actual situation; if the autonomous driving system still has the steering ability, then adopt comfortable steering or emergency steering according to the actual situation.

[0024] The risk mitigation measures corresponding to the fifth safe state slipping path are to enable the autonomous driving redundancy system to take over the vehicle and perform a safe stop.

[0025] Optionally, independent functional logic links and execution links are designed for the five risk mitigation measures of short braking, comfortable braking, emergency braking, comfortable steering and emergency steering, which are arbitrated by the braking behavior arbitration and the steering behavior arbitration respectively, and the final vehicle control instructions are output.

[0026] In a second aspect, a risk mitigation control system for human-machine co-driving according to the present invention includes a memory and a controller, the controller is connected to the memory, and a computer-readable program is stored in the memory. When the controller calls the computer-readable program, it can execute the steps of the risk mitigation control method for human-machine co-driving according to the present invention.

[0027] In a third aspect, a vehicle according to the present invention adopts the risk mitigation control system for human-machine co-driving according to the present invention.

[0028] The present invention has the following advantages: The present invention takes the driver as the safety backup of the autonomous driving system, and on this basis, a general guiding ideology for solving the safety problem of autonomous driving is proposed. The present invention can effectively evaluate the safety of human-machine co-driving and comprehensively give different risk mitigation strategies according to the safety of humans and machines. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a schematic diagram of the human-machine co-driving model in this embodiment;

[0030] Figure 2 is a design flow chart of the risk mitigation strategy in this embodiment;

[0031] Figure 3 is an example diagram of a risk mitigation strategy in this embodiment. Detailed implementation manner

[0032] The present invention will be further described below with reference to the accompanying drawings.

[0033] In this embodiment, a risk mitigation control method for human-machine co-driving includes the following steps:

[0034] Step 1. Build a human-machine co-driving model oriented to safety:

[0035] Take the weakening degree and weakening speed of the driver's takeover ability, and the weakening degree and weakening speed of the automatic driving system control ability as input signals. Take the safety measures corresponding to the safety state sliding path as output signals. Establish a human-machine co-driving model between the input signals and the output signals; wherein, the human-machine co-driving model is configured to: calculate the safety state sliding path according to the weakening degree and weakening speed of the driver's takeover ability, and the weakening degree and weakening speed of the automatic driving system control ability, and output the risk mitigation measure trigger request and response priority corresponding to the safety state sliding path.

[0036] Step 2. Monitor the driver's takeover ability and the automatic driving system control ability in real time and input them into the human-machine co-driving model. The human-machine co-driving model calculates the safety state sliding path according to the input signals, and outputs the risk mitigation measure trigger request and response priority for each safety state sliding path.

[0037] Step 3. Collect the risk mitigation measure trigger requests under all safety state sliding paths, and respond to the risk mitigation measure trigger request with the highest priority.

[0038] As Figure 1 shown, the human-machine co-driving model in this embodiment first defines the driver's state. As a safety backup for vehicle control, the driver's ability to take over the vehicle and safely perform driving actions is the most critical. This human-machine co-driving model takes the driver's takeover ability as a reference dimension. The driver's takeover ability is distributed between "having takeover ability" and "not having takeover ability". Here, "takeover" means that the driver can completely independently control the lateral and longitudinal movements of the vehicle and has the ability to independently judge whether the vehicle is driving safely. The driver's takeover ability is affected by many complex factors, including natural environmental factors, psychological factors, physical health status, mental state, etc. The driver's state may also be unstable or sudden, and the strength of his takeover ability changes in real time with the scene.

[0039] Figure 1Another dimension for evaluating safety risks is also presented, that is, the ability of the autonomous driving system to control the vehicle. The ability of the autonomous driving system to control the vehicle is distributed between "having control ability" and "having no control ability". Among them, "control ability" not only includes the lateral and longitudinal motion control abilities of the vehicle, but also includes the potential hazard handling abilities that may affect vehicle safety, such as human-machine interaction, system fault diagnosis, and fault handling. During the operation of the vehicle, the ability of the autonomous driving system to control the vehicle changes in real time with the traffic scenario and road conditions. Existing autonomous driving systems have the ability to evaluate their own control ability.

[0040] In this embodiment, the situation where the driver has the takeover ability and the autonomous driving system has the control ability is defined as a safe state; the situation where the driver has no takeover ability and the autonomous driving system has the control ability is defined as the first safe state; the situation where the driver has the takeover ability and the autonomous driving system has no control ability is defined as the second safe state; the situation where the driver has no takeover ability and the autonomous driving system has no control ability is defined as an unsafe state. Although the driver's takeover ability and the autonomous driving system's control ability change continuously, for the sake of easy description, as Figure 1 shown, the two-dimensional space in it is divided into four quadrants. The upper left quadrant represents the safe state; the upper right quadrant represents the first safe state, the lower left quadrant represents the second safe state; and the lower right quadrant represents the unsafe state.

[0041] In this embodiment, the transition from the safe state to the first safe state is defined as the first safe state slip path (abbreviated as path 1); the transition from the safe state to the second safe state is defined as the second safe state slip path (abbreviated as path 2); the transition from the first safe state to the unsafe state is defined as the third safe state slip path (abbreviated as path 3); the transition from the second safe state to the unsafe state is defined as the fourth safe state slip path (abbreviated as path 4); the transition from the safe state to the unsafe defined state is defined as the fifth safe state slip path (abbreviated as path 5).

[0042] Figure 1 Five safe state slip paths are marked in it. The reasons for the occurrence, the degree of danger, and the countermeasures of each safe state slip path are different. In the actual scenario, continuous slipping may also occur. In this embodiment, reverse paths are not considered, that is, the paths where the safety level improves from the lower right quadrant to the upper left quadrant, because only risk mitigation schemes are concerned in this embodiment.

[0043] As Figure 1 shown, the combined paths {1, 3} and {2, 4} represent a relatively slow safety deterioration process; path 5 represents a rapid safety deterioration process. The design goal of the risk mitigation strategy is to prevent the human-machine co-driving system from slipping into the lower right quadrant. Therefore, safety measures need to be designed on each slip path.

[0044] In this embodiment, risk mitigation measures need to be specially designed according to different scenarios and system functions. In driving assistance functions and autonomous driving functions, due to the different degrees of driver participation in vehicle control, even for the same dangerous scenario, the risk mitigation measures taken should be different. Risk mitigation measures include various types such as takeover alarms, function degradation, safe parking, emergency braking, and emergency steering. One or more measures can be included in the risk mitigation measures for each safety state sliding path, and it is necessary to combine pre-event prevention and post-event handling. As Figure 2 shown in

[0045] As Figure 3As shown, a risk mitigation strategy based on braking and steering is presented, aiming to gain more time for the driver to take over in the autonomous driving mode. For Path 1 and Path 2, to prevent either the human or the machine from entering an uncontrollable state, the driver's attention should first be drawn through a takeover alarm, and the driver should be informed to immediately take over the vehicle. Short braking is a more intense tactile alarm method than a sound alarm and can be used as a risk mitigation measure for Path 1 and 2 considering the user experience. For Path 3, the driver's intention to take over is still not obvious during this process. Even if the collision risk has not occurred or the dangerous situation is within the driver's handling range, to avoid simultaneous loss of control of the human and the machine, comfortable braking can be used to reduce the vehicle speed, and a more intense takeover alarm should be continuously issued to remind the driver to take over as soon as possible. If an unexpected collision risk or a dangerous situation exceeds the driver's handling range, emergency braking can be used to minimize the collision damage or even avoid the collision. Also combined with the assessment of the collision risk, the system can control the vehicle's steering to keep the vehicle driving within the lane or even avoid obstacles. For Path 4, the system has partially or completely relinquished the vehicle control right (either passively exited due to a fault or actively exited by the system), and the driver should immediately take over. During this process, the system needs to continue to control the vehicle's movement using the remaining control capabilities to avoid a control vacuum caused by the driver's failed takeover. If the system still has the braking ability, like the risk mitigation measure for Path 3, comfortable braking or emergency braking can be used. If the system still has the steering ability, like the risk mitigation measure for Path 3, comfortable braking or emergency braking can be used. Additionally, for Path 3 and 4, after the driver takes over, the system can still provide necessary assistance to the driver, such as providing greater braking force or steering angle when the driver's braking or steering is insufficient. Path 5 corresponds to the most urgent dangerous scenario. At this time, the driver is unable to respond to the danger, the system fails simultaneously, and the vehicle risks falling into a control vacuum. When Path 5 is inevitable, redundant design should be added to the risk mitigation measures, and the autonomous driving redundant system needs to take over the vehicle and perform a safe stop.

[0046] As Figure 3 shown, independent functional logic links and execution links are designed for five risk mitigation measures: short braking, comfortable braking, emergency braking, comfortable steering, and emergency steering. Braking and steering are arbitrated separately and the final vehicle control commands (OUT-1) and (OUT-2) are output. It should be noted especially that Figure 3The (IN-6) contains information on whether the driver has taken over, which can be used as an inhibition condition in each triggering logic to prevent the issuance of braking or steering trigger requests. The (IN-6) should also include information on driver activity, such as the frequency and amplitude of the driver's operations on the accelerator, brake, and steering wheel, etc., for judging the driver's takeover ability. The (IN-5) contains various information on system failures. Different types of system failures have different impacts on the system control ability and need to be specially handled in each risk mitigation measure. The (IN-4) indicates that the lateral, longitudinal, or lateral and longitudinal movement of the vehicle is being controlled by the system. The (F-1) is the degree of collision risk predicted by the environmental perception model.

[0047] In this embodiment, a risk mitigation control system for human-machine co-driving includes a memory and a controller. The controller is connected to the memory. The memory stores a computer-readable program. When the controller calls this computer-readable program, it can execute the steps of the risk mitigation control method for human-machine co-driving as described in this embodiment.

[0048] In this embodiment, a vehicle adopts the risk mitigation control system for human-machine co-driving as described in this embodiment.

Claims

1. A risk mitigation control method for human-machine co-driving, characterized in that, it includes the following steps: Step 1. Build a human-machine co-driving model for safety: Take the weakening degree and weakening speed of the driver's takeover ability, and the weakening degree and weakening speed of the automatic driving system control ability as input signals; Take the safety measures corresponding to the safety state sliding path as output signals; Establish a human-machine co-driving model between the input signal and the output signal; wherein, the human-machine co-driving model is configured to: calculate the safety state sliding path according to the weakening degree and weakening speed of the driver's takeover ability, and the weakening degree and weakening speed of the automatic driving system control ability, and output the risk mitigation measure trigger request and response priority corresponding to the safety state sliding path; Step 2. Real-time monitor the driver's takeover ability and the automatic driving system control ability and input them into the human-machine co-driving model. The human-machine co-driving model calculates the safety state sliding path according to the input signal, and outputs the risk mitigation measure trigger request and response priority for each safety state sliding path; Step 3. Collect the risk mitigation measure trigger requests under all safety state sliding paths, and respond to the risk mitigation measure trigger request with the highest priority; The driver's takeover ability is distributed between "has takeover ability" and "has no takeover ability"; the ability of the automatic driving system to control the vehicle is distributed between "has control ability" and "has no control ability"; Define the situation where the driver has takeover ability and the automatic driving system has control ability as a safe state; Define the situation where the driver has no takeover ability and the automatic driving system has control ability as the first safe state; Define the situation where the driver has takeover ability and the automatic driving system has no control ability as the second safe state; Define the situation where the driver has no takeover ability and the automatic driving system has no control ability as an unsafe state; Define the transition from the safe state to the first safe state as the first safety state sliding path; Define the transition from the safe state to the second safe state as the second safety state sliding path; Define the transition from the first safe state to the unsafe state as the third safety state sliding path; Define the transition from the second safe state to the unsafe state as the fourth safety state sliding path; Define the transition from the safe state to the unsafe defined state as the fifth safety state sliding path.

2. The risk mitigation control method for human-machine co-driving according to claim 1, characterized in that: The risk mitigation measures corresponding to the first safety state sliding path and the second safety state sliding path are to arouse the driver's attention through a takeover alarm method of short braking, and inform the driver to immediately take over the vehicle; The risk mitigation measure corresponding to the third safety state sliding path is to use a comfortable braking method to reduce the vehicle speed, and continuously issue a stronger takeover alarm to remind the driver to take over as soon as possible; if there is an unexpected collision danger or the dangerous situation exceeds the range that the driver can handle, then use an emergency braking method, and the automatic driving system can control the vehicle to steer; The risk mitigation measure corresponding to the fourth safety state sliding path is that the driver is required to immediately complete takeover. During this process, the autonomous driving system needs to continue to control the vehicle movement using the remaining control capabilities. If the autonomous driving system still has braking capabilities, comfort braking or emergency braking shall be adopted according to the actual situation; if the autonomous driving system still has steering capabilities, comfort steering or emergency steering shall be adopted according to the actual situation. The risk mitigation measure corresponding to the fifth safety state sliding path is to enable the autonomous driving redundant system to take over the vehicle and perform a safe stop.

3. The risk mitigation control method for human-machine co-driving according to claim 2, characterized in that: Independent functional logic links and execution links are designed for the five risk mitigation measures of short braking, comfort braking, emergency braking, comfort steering, and emergency steering, and are arbitrated by the braking behavior arbitration and the steering behavior arbitration respectively, and the final vehicle control instruction is output.

4. A risk mitigation control system for human-machine co-driving, comprising a memory and a controller, the controller is connected to the memory, characterized in that: The memory stores a computer-readable program. When the controller calls the computer-readable program, it can execute the steps of the risk mitigation control method for human-machine co-driving according to any one of claims 1 to 3.

5. A vehicle, characterized in that: It adopts the risk mitigation control system for human-machine co-driving according to claim 4.

Images