Emergency man-machine co-driving takeover control method for autonomous vehicle

Abstract

The invention relates to an automatic driving vehicle emergency man-machine co-driving takeover control method, which comprises the following steps: under the condition that a vehicle is in a man-machine co-driving mode, determining a predicted driver intention type based on driver operation data, vehicle external environment data and a preset intention identification model; and determining target control parameters of the vehicle based on a preset driving right distribution strategy, the vehicle operation data and the predicted driver intention type. By adopting the method, the driving safety of automatic driving can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic driving, in particular to an automatic driving vehicle emergency human-machine co-driving takeover control method and device, computer equipment, readable storage medium and program product. BACKGROUND

[0002] With the rapid development of automatic driving technology, a large number of vehicles with Level 2 Partial Driving Automation (L2) and Level 3 Conditional Driving Automation (L3) automatic driving functions have appeared. During the driving process of the automatic driving vehicle, it is inevitable to encounter scenes that cannot be recognized by the partial automatic driving system, such as automatic driving long tail scenes. In this emergency, the L2 and L3 level automatic driving vehicle needs to return the vehicle driving right to the human driver.

[0003] The automatic driving technology in the related art usually directly transfers the control right of the vehicle to the human driver when encountering an emergency and needing to transfer the vehicle driving right to the human driver. However, the human driver often cannot quickly transfer from a non-driving task to a driving task, and it is difficult to quickly obtain the scene awareness of the vehicle state and the surrounding environment, which may lead to incorrect control of the vehicle and cause accidents. SUMMARY

[0004] Therefore, it is necessary to provide an automatic driving vehicle emergency human-machine co-driving takeover control method capable of improving the driving safety of automatic driving in view of the above technical problems.

[0005] In a first aspect, the present application provides an automatic driving vehicle emergency human-machine co-driving takeover control method, comprising:

[0006] In the case that the vehicle is in a human-machine co-driving mode, a predicted driver intention type is determined based on driver operation data, vehicle external environment data and a preset intention recognition model.

[0007] A target control parameter of the vehicle is determined based on a preset driving right allocation strategy, the vehicle operation data and the predicted driver intention type.

[0008] In one of the embodiments, the target control parameter of the automatic driving vehicle is determined based on the preset driving right allocation strategy, the vehicle operation data and the predicted driver intention type, comprising:

[0009] generate a safe driving path of the vehicle based on the predicted driver intention type and a preset path planning strategy, and determine planning operation data corresponding to the safe driving path of the vehicle;

[0010] determine a target control parameter of the vehicle based on the planning operation data, the vehicle operation data, and a preset driving right allocation strategy.

[0011] In one of the embodiments, the vehicle operation data is first front wheel steering angle data of driver operation, the planning operation data is second front wheel steering angle data of vehicle control, and the determination of the target control parameter of the vehicle based on the planning operation data, the vehicle operation data, and the preset driving right allocation strategy comprises:

[0012] determining a driver driving right coefficient and a vehicle driving right coefficient corresponding to the preset driving right allocation strategy;

[0013] determining a product value of the driver driving right coefficient and the first front wheel steering angle data as driver control front wheel steering angle data, and determining a product value of the vehicle driving right coefficient and the second front wheel steering angle data as vehicle control front wheel steering angle data;

[0014] determining a sum value of the driver control front wheel steering angle data and the vehicle control front wheel steering angle data as target front wheel steering angle data of the vehicle.

[0015] In one of the embodiments, the preset intention recognition model at least comprises a double-layer long short-term memory network layer, a first full connection layer, a second full connection layer, and a normalized exponential function layer, and the determination of the predicted driver intention type based on the driver operation data, the vehicle external environment data, and the preset intention recognition model comprises:

[0016] determining vehicle driving data under driver control based on the driver operation data and a vehicle dynamics model;

[0017] inputting the vehicle driving data and the vehicle external environment data into the double-layer long short-term memory network layer for feature extraction processing to obtain a first feature vector;

[0018] inputting the first feature vector into the first full connection layer for feature extraction processing to obtain a second feature vector;

[0019] inputting the second feature vector into the second full connection layer for feature extraction processing to obtain a third feature vector;

[0020] inputting the third feature vector into the normalized exponential function layer for classification processing to obtain probability values corresponding to a plurality of preset driver intention types, and determining a preset driver intention type corresponding to the highest probability value as the predicted driver intention type.

[0021] In one of the embodiments, the method further comprises:

[0022] If the vehicle is in the automatic driving mode and a vehicle abnormal signal is detected, output a request takeover signal and switch to the man-machine co-driving mode; when switching to the man-machine co-driving mode reaches a preset time length, perform man-machine co-driving control on the vehicle based on the target control parameter.

[0023] In one of the embodiments, the method further comprises:

[0024] If no vehicle abnormal signal is detected and a trigger signal of the automatic driving mode is received, switch to the automatic driving mode.

[0025] In one of the embodiments, the vehicle driving data includes lateral driving speed, longitudinal driving speed, lateral driving acceleration and longitudinal driving acceleration of the vehicle, the driver operation data includes steering wheel angle data operated by the driver and pedal opening value operated by the driver, the vehicle dynamics model includes a two-degree-of-freedom vehicle dynamics model and a longitudinal vehicle dynamics model, and the determination of the vehicle driving data under the control of the driver based on the driver operation data and the vehicle dynamics model comprises:

[0026] determining the lateral driving speed based on the two-degree-of-freedom vehicle dynamics model and the steering wheel angle data, and determining the lateral driving acceleration based on the lateral driving speed;

[0027] determining the longitudinal driving speed based on the longitudinal vehicle dynamics model and the pedal opening value, and determining the longitudinal driving acceleration based on the longitudinal driving speed.

[0028] In a second aspect, the present application further provides an automatic driving vehicle emergency man-machine co-driving takeover control device, comprising:

[0029] a first determination module configured to determine a predicted driver intention type based on driver operation data, vehicle external environment data and a preset intention recognition model when the vehicle is in a man-machine co-driving mode;

[0030] a second determination module configured to determine a target control parameter of the vehicle based on a preset driving right allocation strategy, the vehicle operation data and the predicted driver intention type.

[0031] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0032] When the vehicle is in human-machine co-driving mode, the driver's intention type is determined based on driver operation data, vehicle external environment data and a preset intention recognition model.

[0033] Based on the preset driving rights allocation strategy, the vehicle operation data, and the predicted driver intention type, the target control parameters of the vehicle are determined.

[0034] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0035] When the vehicle is in human-machine co-driving mode, the driver's intention type is determined based on driver operation data, vehicle external environment data and a preset intention recognition model.

[0036] Based on the preset driving rights allocation strategy, the vehicle operation data, and the predicted driver intention type, the target control parameters of the vehicle are determined.

[0037] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0038] When the vehicle is in human-machine co-driving mode, the driver's intention type is determined based on driver operation data, vehicle external environment data and a preset intention recognition model.

[0039] Based on the preset driving rights allocation strategy, the vehicle operation data, and the predicted driver intention type, the target control parameters of the vehicle are determined.

[0040] The aforementioned emergency human-machine co-driving takeover control method for autonomous vehicles, while the vehicle is in human-machine co-driving mode, determines the predicted driver intention type based on driver operation data, vehicle external environment data, and a preset intention recognition model. It also determines the vehicle's target control parameters based on a preset driving rights allocation strategy, vehicle operation data, and the predicted driver intention type. Based on these target control parameters, human-machine co-driving steering control is implemented. This ensures smooth vehicle control even when the driver is in a state of loss of awareness during the switch from autonomous driving mode to human-machine co-driving mode, and allows for the transfer of driving rights to the human driver once the driver returns to a normal state. Furthermore, it ensures that in emergency situations, the system assists the human driver in handling dangerous scenarios based on the human driver's evasive intentions, thus avoiding human-machine conflict issues. Attached Figure Description

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the accompanying drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the accompanying drawings in the following description only represent some embodiments of the present application, and for those skilled in the art, other related drawings can also be obtained without creative labor based on these drawings.

[0042] Figure 1 A flowchart of an automatic driving car emergency human-machine co-driving takeover control method in an embodiment;

[0043] Figure 2 A network structure diagram of a preset intention recognition model in an embodiment;

[0044] Figure 3 A flowchart of an automatic driving car emergency human-machine co-driving takeover control method in an embodiment;

[0045] Figure 4 A structural block diagram of an automatic driving car emergency human-machine co-driving takeover control device in an embodiment;

[0046] Figure 5 An internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0048] In an exemplary embodiment, as shown in Figure 1 An automatic driving car emergency human-machine co-driving takeover control method is provided, and the present embodiment is exemplified by the method applied in the human-machine co-driving system of the vehicle. In the present embodiment, the method comprises the following steps:

[0049] Step 101, in the case that the vehicle is in a human-machine co-driving mode, determining a predicted driver intention type based on driver operation data, vehicle external environment data and a preset intention recognition model.

[0050] The human-machine co-driving mode can refer to a driving mode in which the intelligent system of the vehicle cooperates with the human driver to jointly complete vehicle control. The driver operation data can represent the operation data of the driver on the vehicle in the human-machine co-driving mode. The vehicle external environment data can include the distance between the ego vehicle and surrounding vehicles, and the surrounding vehicles can include vehicles in front, on the left, on the right, and behind the ego vehicle. The preset intention recognition model can be used to identify the avoidance intention of the driver to the obstacle. The preset intention recognition model can be a pre-trained neural network, which can be a Long Short-Term Memory (LSTM) network. The driver intention type can be one or more of left avoidance, right avoidance, and straight driving.

[0051] Specifically, when the vehicle is in an automatic driving mode, the control system of the vehicle can obtain sensor data of each part of the vehicle, which can include vehicle internal driving data (for example: lateral position data, longitudinal position data, lateral speed data, longitudinal position data, yaw angle data of the ego vehicle, driver steering operation data, etc.) and vehicle surrounding environment data (for example: lateral position data, longitudinal position data, lateral speed data, longitudinal position data of vehicles around the vehicle, or lateral position data, longitudinal position data of obstacles around the vehicle, etc.). When the control system of the vehicle detects an abnormal signal of the vehicle, a request takeover signal is output, which is used to remind the human driver that the current vehicle is in an emergency state and to remind the human driver to take over the steering control of the vehicle. Alternatively, the abnormal signal can be an abnormal signal generated by the ego vehicle when the sensor detection data of the ego vehicle is abnormal.

[0052] Upon receiving the request takeover signal, the vehicle starts a human-machine co-driving mode. The human-machine co-driving system can determine the longitudinal position data of the surrounding vehicles in the current scenario, determine the distance between the ego vehicle and the surrounding vehicles based on the longitudinal position data, and determine the distance as the vehicle external environment data. The human-machine co-driving system can determine the vehicle driving data under the virtual operation of the driver according to the driver operation data. The human-machine co-driving system can input the vehicle external environment data and the vehicle driving data into the preset intention recognition model to obtain the predicted driver intention type.

[0053] Step 102, based on the preset driving right allocation strategy, the vehicle operation data and the predicted driver intention type, determining the target control parameter of the vehicle.

[0054] The vehicle operation data can be operation data of a steering component in the vehicle when the driver controls the steering of the vehicle during driving of the vehicle, such as a steering angle of the steering component controlled by the driver, and the like. The preset driving right allocation strategy can be used to determine the driving right allocation coefficient of the human driver and the vehicle driving right allocation coefficient. The target control parameter is a control parameter that is jointly operated by the human driver and the automatic driving system of the vehicle on the steering component of the vehicle.

[0055] Specifically, the man-machine co-driving system can determine corresponding vehicle operation data based on the driver operation data. The man-machine co-driving system can determine a vehicle safe driving path based on the predicted driver intention type, and determine planning operation data corresponding to the vehicle safe driving path; and determine the target control parameter of the vehicle based on the preset driving right allocation strategy, the first front wheel steering angle data, and the planning operation data.

[0056] The above-mentioned automatic driving vehicle emergency man-machine co-driving takeover control method, by determining the predicted driver intention type based on the driver operation data, the vehicle external environment data and the preset intention recognition model when the vehicle is in the man-machine co-driving mode, and determining the target control parameter of the vehicle based on the preset driving right allocation strategy, the vehicle operation data and the predicted driver intention type. Based on the target control parameter, the man-machine co-driving steering control of the vehicle is realized, which realizes that when switching from the automatic driving mode to the man-machine co-driving mode, even if the driver is in a situational loss state, the vehicle control right can be smoothly guaranteed and the driving right can be transferred to the human driver when the driver recovers to the normal state, and at the same time, the system can assist the human driver to handle the dangerous scene according to the avoidance intention of the human driver when the automatic driving vehicle encounters an emergency, and the problem of man-machine conflict is avoided.

[0057] In an exemplary embodiment, the specific implementation process of the step 102 "determining the target control parameter of the automatic driving vehicle based on the preset driving right allocation strategy, the vehicle operation data and the predicted driver intention type" can include:

[0058] Based on the predicted driver intention type and the preset path planning strategy, a vehicle safe driving path is generated, and planning operation data corresponding to the vehicle safe driving path is determined; based on the planning operation data, the vehicle operation data and the preset driving right allocation strategy, the target control parameter of the vehicle is determined.

[0059] The vehicle safe driving path is a predicted driving path of the vehicle under the current environment. The planning operation data is data of the automatic driving system of the vehicle operating the steering component of the vehicle during driving of the vehicle safe driving path, such as steering wheel steering angle data, front wheel steering angle data, and the like.

[0060] Specifically, the man-machine co-driving system can input the preset driver intention type into the preset planning path strategy for processing to obtain a vehicle safe driving path. The man-machine co-driving system can identify path data of the vehicle safe driving path, determine corresponding planning operation data of the vehicle based on the path data and vehicle dynamics principles. The man-machine co-driving system can determine a vehicle driving right coefficient corresponding to the planning operation data and a driver driving right coefficient corresponding to the vehicle operation data based on a preset driving right allocation strategy, and determine a target control parameter of the vehicle based on the vehicle driving right coefficient, the planning operation data, the vehicle operation data, and the driver driving right coefficient.

[0061] Optionally, the preset path planning strategy can be based on a cubic polynomial fitting method or a B-spline interpolation method, etc., which is not limited here. When the preset path planning strategy is based on a cubic polynomial fitting method, the method for determining the vehicle safe driving path can include: taking the current position of the vehicle as the origin, determining the driving direction of the vehicle as the y-axis, determining the lateral offset direction of the vehicle as the x-axis, and establishing a coordinate system; determining an initial cubic polynomial fitting model, determining the driving endpoint position coordinates and the driving direction angle value of the vehicle based on the preset driver intention type, and determining the vehicle safe driving path based on the driving position coordinates, the direction angle value, and the initial cubic polynomial fitting model.

[0062] When the preset path planning strategy is a B-spline interpolation method, the method for determining the vehicle safe driving path can include the following steps: determining the path key points corresponding to the driver intention type based on the preset driver intention type, taking the path key points as the control points of the B-spline, and determining a preset order, calculating the B-spline basis function based on the control points and the preset order, generating an initial path curve through a difference value generation algorithm and the B-spline basis function, and optimizing the initial path curve based on the external environment data of the vehicle to obtain the vehicle safe driving path.

[0063] In this embodiment, the target control parameter is used to realize man-machine co-driving steering control of the vehicle, which realizes smooth transfer of vehicle control right when switching from an automatic driving mode to a man-machine co-driving mode, even if the driver is in a situational loss state, and ensures that the driving right is transferred to the human driver when the driver recovers to a normal state.

[0064] In an exemplary embodiment, the specific implementation process of the step "determining the target control parameter of the vehicle based on the planning operation data, the vehicle operation data, and the preset driving right allocation strategy" can include:

[0065] determining a driver driving right coefficient and a vehicle driving right coefficient corresponding to the preset driving right allocation strategy; determining a product value of the driver driving right coefficient and the first front wheel steering angle data as driver-controlled front wheel steering angle data; determining a product value of the vehicle driving right coefficient and the second front wheel steering angle data as vehicle-controlled front wheel steering angle data; and determining a sum value of the driver-controlled front wheel steering angle data and the vehicle-controlled front wheel steering angle data as target front wheel steering angle data of the vehicle.

[0066] wherein the vehicle operation data is the first front wheel steering angle data operated by the driver. The planned operation data is the second front wheel steering angle data controlled by the vehicle. The driver driving right coefficient is the proportion of driving right allocated to the driver by the man-machine co-driving system in the process of man-machine co-driving. The vehicle driving right coefficient is the proportion of driving right allocated to the vehicle by the man-machine co-driving system in the process of man-machine co-driving. The driver-controlled front wheel steering angle data is the data of the front wheel steering angle that can be controlled by the human driver. The vehicle-controlled front wheel steering angle data is the data of the front wheel steering angle controlled by the vehicle system. The target front wheel steering angle data is the data of the controlled front wheel steering angle finally output by the vehicle.

[0067] Specifically, the man-machine co-driving system can determine a difference value between the first longitudinal position data of the vehicle front object and the second longitudinal position data of the vehicle as position difference data; and determine a difference value between the second longitudinal speed data of the vehicle and the first longitudinal speed data of the vehicle front object as speed difference data; and determine a ratio of the position difference data and the speed difference data as the Time-to-Collision (TTC) of the ego vehicle. The specific calculation formula of the Time-to-Collision can be:

[0068]

[0069] wherein TTC is the Time-to-Collision, x p is the first longitudinal position data of the vehicle front object, x e is the second longitudinal position data of the ego vehicle. is the second longitudinal speed data of the vehicle. is the first longitudinal speed data of the vehicle front object.

[0070] The man-machine co-driving system can determine a ratio of the position difference data and the second longitudinal speed data as the Time Headway (TH) of the ego vehicle. The specific calculation formula of the Time Headway can be:

[0071]

[0072] wherein TH is the Time Headway.

[0073] The man-machine co-driving system can determine a first threshold value and a first ratio of the headway, determine a second threshold value, a product value of the gravitational acceleration and the collision time, and determine a second ratio of the speed difference data and the product value, and determine a sum value of the first ratio and the second ratio as the environmental risk evaluation index of the vehicle in the target state. The first threshold value can be determined according to the reaction time of the driver, for example, the first threshold value can be 1, and the second threshold value can be determined according to the limit deceleration of the vehicle, for example, the second threshold value can be 1.5. The specific calculation formula of the environmental risk evaluation index can be:

[0074]

[0075] wherein EH is the environmental risk evaluation index, and g is the gravitational acceleration of the current location.

[0076] The vehicle controller can generate prediction path data based on the current environment. The man-machine co-driving system can determine the prediction lateral position data of the vehicle based on the current longitudinal position data and the prediction path data of the vehicle. The man-machine co-driving system can take the derivative of the prediction path data, input the longitudinal position data into the derivative, and determine the prediction yaw angle data of the vehicle. The man-machine co-driving system can determine the difference between the lateral position data of the ego vehicle and the prediction lateral position data as the lateral position tracking error of the vehicle, and determine the difference between the yaw angle data and the prediction yaw angle data as the yaw angle tracking error of the vehicle. The calculation formula of the lateral position tracking error and the yaw angle tracking error can be:

[0077]

[0078]

[0079] wherein e y is the lateral displacement tracking error of the vehicle, is the yaw angle tracking error of the vehicle. y e is the longitudinal position data, f(x) is the prediction path data, f(x e ) is the prediction lateral position data, is the lateral position data of the ego vehicle, is the derivative of the prediction path data, is the prediction yaw angle data.

[0080] The man-machine co-driving system can calculate a first square value of the product value of the first constant and the lateral displacement tracking error, calculate a second square value of the product value of the second constant and the yaw angle tracking error, calculate a sum value of 1, the first square value and the second square value, and determine the reciprocal of the sum value as the man-machine driving target correlation index (Target Correlation, TC). The specific calculation formula of the man-machine driving target correlation index can be:

[0081]

[0082] wherein, TC is the human-machine driving target correlation evaluation index, a1 is a first constant, a2 is a second constant, the first constant and the second constant are used to adjust the relative weight of the vehicle lateral position tracking error and the yaw angle tracking error, for example, a1 = 1, a2 = 0.2, it should be understood that the first constant and the second constant are only used for example and do not constitute a specific limitation, and specific data can be determined according to specific application scenarios.

[0083] The human-machine co-driving system can determine the sum value of the square value of the front wheel steering angle data in the preset time period before the target time, determine the sum value of the square value of the front wheel steering angle rate in the preset time period before the target time, determine the fourth power value of the third constant of the above two sum values, calculate the product value of the fourth power value and the fifth constant, and determine the sum value of the product value and the fifth constant as the driver takeover intention evaluation index (TA). The specific calculation formula of the driver takeover intention evaluation index can be:

[0084]

[0085] wherein, TA is the human driver takeover intention evaluation index, t0 is the current time, is the front wheel steering angle data at time t, is the front wheel steering angle rate at time t, β1 is the fourth constant, β2 is the third constant, β3 is the fifth constant, β1, β2 and β3 are used to adjust the size of the human driver takeover intention evaluation index. β1, β2 and β3 are obtained by fitting through driver bench test and real feeling evaluation, for example, β1 = 0.9, β2 = 0.14, β3 = 0.57, it should be understood that the examples here are only used for example and cannot constitute a specific limitation.

[0086] The human-machine co-driving system can calculate the seventh power value of the product value of the sixth constant and the human-machine driving target correlation evaluation index, calculate the ninth power value of the product value of the eighth constant and the driver takeover intention evaluation index, calculate the product value of the seventh power value and the ninth power value, calculate the negative exponential function value of the product value with the base of natural constant e, and determine the difference between 1 and the negative exponential function value as the driver system adaptation evaluation index (DA). The specific calculation formula of the driver system adaptation evaluation index can be:

[0087]

[0088] Wherein, DA is the driver system adaptability evaluation index, γ1 is the sixth constant, γ2 is the seventh constant, γ3 is the eighth constant, γ4 is the ninth constant, which is obtained by parameter adjustment, γ1, γ2, γ3, γ4 are used to adjust the influence of the human-machine driving target correlation and the willingness of the human driver to take over on the driver system adaptability, which can be determined through driver bench test, for example, γ1=2, γ2=1, γ3=3, γ4=3, it should be understood that the examples here are only for example and cannot constitute specific limitation.

[0089] The human-machine co-driving system can determine a first product value of the first coefficient and the environmental risk evaluation index, and a first difference value of the preset value and the first product value; determine a second product value of the second coefficient and the environmental risk evaluation index; determine a third product value of the third coefficient and the driver system adaptability evaluation index, determine a first sum value of the target moment and the third product value, determine a second difference value of the first sum value and the fourth coefficient, determine a fourth product value of the driver system adaptability evaluation index based on the preset exponential decay strategy, obtain an exponential function value by processing the fourth product value; and determine a second sum value of the exponential function value and the preset value; determine the ratio of the first difference value and the second sum value; and determine the sum value of the ratio and the second product value as the driver driving right coefficient.

[0090] Taking the time when the vehicle system discovers the obstacle and sends a request to the human driver to take over as the starting time, for the human-machine co-driving system, the change rule of the driver driving right coefficient can be expressed as:

[0091]

[0092] Wherein, is the driver driving right coefficient in the human-machine co-driving system, t is the time relative to the starting time, λ1 is the first coefficient, λ2 is the second coefficient, λ3 is the third coefficient, λ4 is the fourth coefficient, which is obtained by parameter adjustment, λ1, λ2, λ3, λ4 are used to adjust the change relationship of the driver driving right coefficient with the system environmental risk and the driver system adaptability, which can be obtained by fitting through driver bench test and real feeling evaluation. For example, λ1=0.75, λ2=0.75, λ3=10, λ4=13, it should be understood that the examples here are only for example and cannot constitute specific limitation.

[0093] The human-machine co-driving system can calculate first front wheel steering angle data applied by the driver of the ego vehicle based on steering wheel steering operation data of the driver steering operation data. The human-machine co-driving system can determine a difference between 1 and the driver driving right coefficient as the vehicle driving right coefficient. The human-machine co-driving system can determine a product value of the driver driving right coefficient and the first front wheel steering angle data applied by the driver of the ego vehicle as the driver control front wheel steering angle data. The human-machine co-driving system can determine a product value of the vehicle driving right coefficient and second front wheel steering angle data of the vehicle based on the safe driving path of the vehicle as the vehicle control front wheel steering angle data. The human-machine co-driving system can determine a sum value of the driver control front wheel steering angle data and the vehicle control front wheel steering angle data as the target front wheel steering angle data. The specific calculation formula of the target front wheel steering angle data can be:

[0094]

[0095] wherein, is the target front wheel steering angle data finally input to the vehicle steering system. is the driver driving right coefficient, is the first front wheel steering angle data caused by the driver's operation on the steering wheel, is the second front wheel steering angle data calculated by the automated system according to the planned path.

[0096] Optionally, the human-machine co-driving system can calculate the steering wheel steering angle data in the driver operation data based on a preset steering angle calculation strategy to obtain the first front wheel steering angle data. The preset steering angle calculation strategy can be steering transmission ratio, nonlinear calibration, chassis test, etc., which are not limited here.

[0097] In the embodiment, the first front wheel steering angle data and the second front wheel steering angle data are dynamically allocated by the driver driving right coefficient and the vehicle driving right coefficient, the target front wheel steering angle data of the vehicle at different times is output, the human-machine co-driving steering control of the vehicle is realized, the vehicle is controlled based on the driver driving right coefficient at the target time, and it is ensured that the driving right can be smoothly transferred to the human driver in time when the automated driving system encounters a situation that needs to be taken over by the human driver, so as to avoid the unsafe problem caused by the loss of situational awareness of the vehicle state and the surrounding environment of the driver due to the non-driving task of the driver, and the safety of the vehicle in the driving process is improved.

[0098] In an exemplary embodiment, the specific implementation process of step 101 "determining the predicted driver intention type based on the driver operation data, the vehicle external environment data, and the preset intention recognition model" can include:

[0099] The vehicle driving data under the control of the driver is determined based on the driver operation data and a vehicle dynamics model; the vehicle driving data and vehicle external environment data are input into a double-layer long short-term memory network layer for feature extraction processing to obtain a first feature vector; the first feature vector is input into a first fully connected layer for feature extraction processing to obtain a second feature vector; the second feature vector is input into a second fully connected layer for feature extraction processing to obtain a third feature vector; the third feature vector is input into a normalized exponential function layer for classification processing to obtain probability values corresponding to a plurality of preset driver intention types, and a preset driver intention type corresponding to the highest probability value is determined as a predicted driver intention type.

[0100] The vehicle driving data is virtual driving data of the vehicle under the control of the driver. The driving data can include longitudinal driving speed, longitudinal driving acceleration, lateral driving speed and lateral driving acceleration. The preset intention recognition model at least includes a double-layer long short-term memory network layer, a first fully connected layer, a second fully connected layer and a normalized exponential function layer. The double-layer long short-term memory network layer can include a first long short-term memory network layer and a second long short-term memory network layer. For example, the long short-term memory network layer can be an LSTM (Long Short-Term Memory) layer. The preset driver intention type can include left avoidance, continue straight and right avoidance.

[0101] Specifically, the man-machine co-driving system can determine the vehicle driving data under the control of the driver based on the driver operation data and the vehicle dynamics model. The man-machine co-driving system can input the driver operation data and the vehicle external environment information as the input of the driver intention recognition network, which can be represented as:

[0102]

[0103] wherein I t is the feature vector of the neural network input at time t. v x is the longitudinal driving speed of the vehicle under the virtual control of the driver, v y is the lateral driving speed of the vehicle under the virtual control of the driver. a x is the longitudinal driving acceleration of the vehicle under the virtual control of the driver; a y is the lateral driving acceleration of the vehicle under the virtual control of the driver, are the longitudinal distances between the ego vehicle and the vehicles in front, left, rear, right, left and right of the ego vehicle, respectively.

[0104] Through the preset intention recognition model, the avoidance intention of the driver at this time can be recognized to obtain the predicted driver intention type. The output of the preset intention recognition model can be represented as:

[0105]

[0106] Among them, O t Let p be the preset intent recognition model of the neural network at time t. l p s p r These represent the probabilities, as determined by the neural network, of whether a driver chooses to swerve to the left, continue straight, or swerve to the right in the face of an emergency.

[0107] Optionally, the preset intent recognition model is constructed based on a two-layer LSTM neural network, and the network structure diagram is as follows: Figure 2 As shown in the figure (where is the tensor dimension; ( ) represents the number of neurons). At each time step, the driver's operation data and the vehicle's external environment data from the past 50 frames are input into a two-layer LSTM neural network. Both the first and second LSTM layers contain 256 neurons. The past 50 frames of data from the current time step ( The features are then fed into the 256 neurons of the first LSTM layer for feature extraction, resulting in an initial feature vector of 256 dimensions. This initial feature vector is then fed into the 256 neurons of the second LSTM layer for feature extraction, resulting in a first feature vector of 256 dimensions. Each of the first feature vector dimensions is then fed into the 32 neurons of the first fully connected layer for feature extraction, resulting in a second feature vector of 32 dimensions. Each of the second feature vector dimensions is then fed into the 3 neurons of the second fully connected layer for feature extraction, resulting in a third feature vector of 3 dimensions. This third feature vector is then fed into a normalized exponential function layer for classification, resulting in probability values ​​corresponding to multiple preset driver intention types. The preset driver intention type corresponding to the highest probability value is then determined as the predicted driver intention type.

[0108] The normalized exponential function layer can be a Softmax function. All learnable layers not mentioned in the network use the Rectified Linear Unit (ReLU) activation function. A dropout layer is added after each fully connected layer to avoid overfitting.

[0109] In the process of neural network training, an adaptive moment estimation (Adam) optimizer is used and an initial learning rate is set to 0.001. Meanwhile, the learning rate will be multiplied by a factor of 0.15 every 200 training cycles. Training is stopped after 1000 cycles are reached. Finally, the preset intention recognition model obtained by training has an identification accuracy of 95.89% on the highD dataset of German vehicle driving data.

[0110] In this embodiment, by using the neural network to identify the driver's intention in an emergency, the probabilities of the driver choosing to dodge left, continue straight and dodge right in this situation are obtained, and the selection with the highest probability is selected as the identified driver's intention. A deep neural network is used to process the driver's driving intention recognition problem, and a double-layer LSTM neural network is used to explore the features in the sequence data. This method is suitable for the prediction problem of automatic driving of such time series data.

[0111] In an exemplary embodiment, the automatic driving vehicle emergency human-machine co-driving takeover control method further comprises:

[0112] If the vehicle is in an automatic driving mode and an abnormal vehicle signal is detected, a request takeover signal is output, and the mode is switched to a human-machine co-driving mode; when the mode is switched to the human-machine co-driving mode for a preset time period, the vehicle is controlled in the human-machine co-driving mode based on the target control parameters.

[0113] Specifically, when the vehicle is in an automatic driving mode, the human-machine co-driving system can obtain vehicle self-vehicle front object state data, self-vehicle surrounding obstacle state data, and self-vehicle driving state data from various sensors in the vehicle in real time during driving. The self-vehicle front state data can also include lateral position data, lateral velocity data, longitudinal position data, and longitudinal velocity data of the front object. The self-vehicle driving state data can also include lateral position data, longitudinal position data, lateral velocity data, longitudinal position data, yaw angle data of the self-vehicle, and driver steering operation data. The terminal can determine the driving mode of the vehicle based on the above data when the vehicle is in a driving state. For example, the driving mode can include an automatic driving mode, a human-machine co-driving mode, a manual driving mode, and the like. In addition, the vehicle can also switch from the automatic driving mode to the human-machine co-driving mode based on the received driving mode switching signal.

[0114] If the vehicle detects that the above data exceeds the normal operation threshold, an abnormal signal is output, and a request takeover signal is output, switching from the autonomous driving mode to the man-machine co-driving mode. The man-machine co-driving system can determine the target control parameter. Within a preset time length of the takeover request signal to the driver, the man-machine co-driving system performs man-machine co-driving control on the vehicle based on the target control parameter. For example, the preset time length can be 0.1s. Alternatively, the abnormal signal can include: sensor data abnormality, control system failure, environment exceeding the autonomous driving capability range (such as extreme weather), vehicle dynamics abnormality (such as sudden skid), etc.

[0115] In addition, the vehicle system detects an abnormal signal, and the abnormal signal indicates that the vehicle is in a long-tail scenario that cannot be handled, and outputs a request takeover request, which is used to remind the driver to take over the vehicle control as soon as possible. For example, it can be reminded by means of instrument warning light flickering and alarm sound.

[0116] In this embodiment, through abnormality detection and gradual takeover, the risk of the autonomous driving system in abnormal situations is reduced, and smooth mode switching and cooperative control reduce the pressure on the driver and improve comfort.

[0117] In an exemplary embodiment, the autonomous vehicle emergency man-machine co-driving takeover control method further comprises:

[0118] If no vehicle abnormal signal is detected, and a trigger signal of the autonomous driving mode is received, the autonomous driving mode is switched to.

[0119] The trigger signal can be a signal that the driver actively activates the autonomous driving, for example, pressing the autonomous driving button, etc.

[0120] Specifically, if no vehicle abnormal signal is detected by the vehicle, and a trigger signal of the autonomous driving mode is received, the autonomous driving mode is switched from the man-machine co-driving mode or the manual driving mode.

[0121] In this embodiment, by the absence of abnormalities and active triggering, it is ensured that the system only enters the autonomous driving when the state is healthy, avoiding potential risks.

[0122] In an exemplary embodiment, the specific implementation process of the step of "determining the vehicle driving data under the control of the driver based on the driver operation data and the vehicle dynamics model" can include:

[0123] Based on the two-degree-of-freedom vehicle dynamics model and the steering wheel angle data, the lateral driving speed is determined, and the lateral driving acceleration is determined based on the lateral driving speed; based on the vehicle longitudinal dynamics model and the pedal opening value, the longitudinal driving speed is determined, and the longitudinal driving acceleration is determined based on the longitudinal driving speed.

[0124] The vehicle driving data includes lateral driving speed, longitudinal driving speed, lateral driving acceleration and longitudinal driving acceleration of the vehicle, the driver operation data includes steering wheel angle data and pedal opening value of the driver operation, and the vehicle dynamics model includes a two-degree-of-freedom dynamics model and a longitudinal dynamics model of the vehicle. The lateral driving speed, the longitudinal driving speed, the lateral driving acceleration and the longitudinal driving acceleration are all vehicle driving data under the virtual control of the driver.

[0125] Specifically, the man-machine co-driving system obtains the initial lateral speed and the initial longitudinal speed of the vehicle, determines the ratio of the initial longitudinal speed and the initial lateral speed, determines the difference between the steering wheel angle data and the ratio as the front wheel side slip angle, determines the negative value of the ratio as the rear wheel side slip angle, determines the product of the side slip stiffness of the front wheel and the front wheel side slip angle as the front tire side slip force, determines the product of the side slip stiffness of the rear wheel and the rear wheel side slip angle as the rear tire side slip force, calculates the lateral driving speed of the vehicle based on the product of the side slip stiffness of the front wheel and the rear wheel side slip stiffness as the rear tire side slip force, and determines the lateral driving acceleration based on the lateral driving speed. The specific calculation formula of the vehicle lateral dynamics formula can be:

[0126]

[0127] wherein m is the mass data of the vehicle, v y is the lateral driving speed of the vehicle, C f is the side slip stiffness of the front wheel, C r is the side slip stiffness of the rear wheel, a f is the front wheel side slip angle, a r is the rear wheel side slip angle.

[0128] The product of the pedal opening value and the maximum torque of the engine is determined as the torque value of the engine, the efficiency of the engine, the torque value and the angular velocity are determined as the output efficiency of the engine, the ratio of the output efficiency and the initial longitudinal speed is determined as the driving force of the vehicle, the longitudinal dynamics equation of the longitudinal driving speed of the vehicle is established based on the mass data of the vehicle, the vehicle driving force, the rolling resistance and the air resistance, the longitudinal driving speed of the vehicle is calculated based on the longitudinal dynamics equation, and the longitudinal driving acceleration is determined based on the change speed of the longitudinal driving speed.

[0129] In an exemplary embodiment, as Figure 3 shown, the automatic driving vehicle emergency man-machine co-driving takeover control method can specifically include the following steps:

[0130] Step 301, the vehicle automatic driving system is in a working state, and the environmental perception system continuously detects the surrounding environment of the vehicle to obtain the position and speed information of the surrounding obstacles or the vehicle.

[0131] Step 302, the vehicle system detects an automatic driving long tail scene that the system cannot handle, and sends a takeover signal to the driver.

[0132] Step 303, the man-machine co-driving takeover system obtains driver operation data related to the driver's takeover intention and vehicle external environment data, which can include the longitudinal and lateral speed and acceleration of the vehicle at this time, and the longitudinal distance between the vehicle and the vehicles in front, left, right and rear of the vehicle.

[0133] Step 304, the driver operation information and vehicle external environment information are taken as input values of a preset intention recognition model, and a preset driver intention type is output.

[0134] Step 305, based on the preset driver intention type and the preset path planning strategy, a safe driving path of the vehicle is generated.

[0135] Step 306, based on the preset driving right allocation strategy, vehicle operation data and vehicle safe driving path, the target control parameters of the vehicle are determined, and the steering control of the vehicle is performed through the target control parameters.

[0136] Step 307, it is judged whether the emergency of the vehicle is resolved, if not, step 306 is continued to be executed; if yes, step 308 is executed.

[0137] Step 308, it is judged whether a trigger signal of manually restarting the automatic driving function by the driver is received, if not, it is continuously judged whether the trigger signal is received; if yes, step 309 is executed.

[0138] Step 309, the vehicle returns to the automatic driving state.

[0139] It should be understood that although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise stated herein, the execution of these steps has no strict sequence limitation, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be alternately executed with other steps or steps or stages in other steps.

[0140] Based on the same inventive concept, the embodiments of the present application also provide an automatic driving vehicle emergency situation man-machine co-driving interface control device for implementing the automatic driving vehicle emergency situation man-machine co-driving interface control method described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more automatic driving vehicle emergency situation man-machine co-driving interface control device embodiments provided below can be referred to the limitations of the automatic driving vehicle emergency situation man-machine co-driving interface control method described above, which will not be described here.

[0141] In one exemplary embodiment, as shown in Figure 4 An automatic driving vehicle emergency situation man-machine co-driving interface control device 40 is provided, comprising: a first determination module 41 and a second determination module 42, wherein:

[0142] The first determination module 41 is configured to determine a predicted driver intention type based on driver operation data, vehicle external environment data, and a preset intention recognition model when the vehicle is in a man-machine co-driving mode.

[0143] The second determination module 42 is configured to determine a target control parameter of the vehicle based on a preset driving right allocation strategy, the vehicle operation data, and the predicted driver intention type.

[0144] In one embodiment, the second determination module 42 is configured to generate a vehicle safe driving path based on the predicted driver intention type and a preset path planning strategy, and determine planning operation data corresponding to the vehicle safe driving path.

[0145] The target control parameter of the vehicle is determined based on the planning operation data, the vehicle operation data, and the preset driving right allocation strategy.

[0146] In one embodiment, the second determination module 42 is configured to determine a driver driving right coefficient and a vehicle driving right coefficient corresponding to the preset driving right allocation strategy.

[0147] The product value of the driver driving right coefficient and the first front wheel steering angle data is determined as driver control front wheel steering angle data, and the product value of the vehicle driving right coefficient and the second front wheel steering angle data is determined as vehicle control front wheel steering angle data.

[0148] The sum of the driver control front wheel steering angle data and the vehicle control front wheel steering angle data is determined as the target front wheel steering angle data of the vehicle.

[0149] In one embodiment, the first determination module 41 is configured to determine vehicle driving data under driver control based on the driver operation data and a vehicle dynamics model.

[0150] The vehicle driving data and vehicle external environment data are input into the bi-layer long short-term memory network layer for feature extraction processing to obtain a first feature vector;

[0151] The first feature vector is input into the first full connection layer for feature extraction processing to obtain a second feature vector;

[0152] The second feature vector is input into the second full connection layer for feature extraction processing to obtain a third feature vector;

[0153] The third feature vector is input into the normalization exponential function layer for classification processing to obtain probability values corresponding to a plurality of preset driver intention types, and a preset driver intention type corresponding to the highest probability value is determined as a predicted driver intention type.

[0154] In one of the embodiments, the first determination module 41 is further configured to output a request takeover signal and switch to the man-machine co-driving mode if the vehicle is in the automatic driving mode and a vehicle abnormal signal is detected, and perform man-machine co-driving control on the vehicle based on the target control parameter when the switch to the man-machine co-driving mode reaches a preset time length.

[0155] In one of the embodiments, the second determination module 42 is further configured to switch to the automatic driving mode if no vehicle abnormal signal is detected and a trigger signal of the automatic driving mode is received.

[0156] In one of the embodiments, the first determination module 41 is configured to determine the lateral driving speed based on the two-degree-of-freedom dynamics model of the vehicle and the steering wheel angle data, and determine the lateral driving acceleration based on the lateral driving speed.

[0157] Determine the longitudinal driving speed based on the longitudinal dynamics model of the vehicle and the pedal opening value, and determine the longitudinal driving acceleration based on the longitudinal driving speed.

[0158] The above-mentioned various modules in the automatic driving vehicle emergency man-machine co-driving takeover control device can be all or partially realized by software, hardware and combinations thereof. The above-mentioned various modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned various modules.

[0159] In one exemplary embodiment, a computer device, which can be a terminal, is provided, and an internal structure diagram of the computer device can be as shown in Figure 5As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be realized through WIFI, mobile cellular network, near field communication (NFC) or other technologies. The computer program is executed by the processor to realize an automatic driving car emergency man-machine co-driving control method. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0160] Those skilled in the art can understand that, Figure 5 The skilled in the art can understand that,

[0161] In one embodiment, a computer device is provided, including a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize the steps in the above method embodiments.

[0162] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to realize the steps in the above method embodiments.

[0163] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by a processor to realize the steps in the above method embodiments.

[0164] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0165] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments of each method. In the embodiments provided in the present application, any reference to memory, database or other medium can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., without being limited thereto.

[0166] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described above, however, any combination of these technical features is deemed to be within the scope of the present application.

[0167] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for ordinary skilled persons in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for emergency human-machine co-driving takeover control of an autonomous vehicle, characterized in that, The method includes: When the vehicle is in human-machine co-driving mode, the driver's intention type is determined based on driver operation data, vehicle external environment data and a preset intention recognition model. Based on the preset driving rights allocation strategy, vehicle operation data, and the predicted driver intention type, the target control parameters of the vehicle are determined.

2. The method according to claim 1, characterized in that, The process of determining the target control parameters of the autonomous vehicle based on a preset driving rights allocation strategy, vehicle operation data, and the predicted driver intention type includes: Based on the predicted driver intention type and the preset route planning strategy, a safe driving route for the vehicle is generated, and the planning operation data corresponding to the safe driving route for the vehicle is determined. Based on the planned operation data, the vehicle operation data, and the preset driving rights allocation strategy, the target control parameters of the vehicle are determined.

3. The method according to claim 2, characterized in that, The vehicle operation data refers to the first front wheel steering angle data operated by the driver, and the planned operation data refers to the second front wheel steering angle data controlled by the vehicle. The determination of the vehicle's target control parameters based on the planned operation data, the vehicle operation data, and a preset driving authority allocation strategy includes: Determine the driver's driving right coefficient and vehicle driving right coefficient corresponding to the preset driving right allocation strategy; The product of the driver's driving rights coefficient and the first front wheel steering angle data is determined as the driver's control front wheel steering angle data; and the product of the vehicle's driving rights coefficient and the second front wheel steering angle data is determined as the vehicle's control front wheel steering angle data; The sum of the driver-controlled front wheel steering angle data and the vehicle-controlled front wheel steering angle data is determined to be the target front wheel steering angle data of the vehicle.

4. The method according to claim 1, characterized in that, The preset intent recognition model includes at least a two-layer long short-term memory network layer, a first fully connected layer, a second fully connected layer, and a normalized exponential function layer. The step of determining the predicted driver intent type based on driver operation data, vehicle external environment data, and the preset intent recognition model includes: Based on the driver's operation data and the vehicle dynamics model, the vehicle driving data under the driver's control is determined; The vehicle driving data and the vehicle external environment data are input into the two-layer long short-term memory network layer for feature extraction processing to obtain the first feature vector. The first feature vector is input into the first fully connected layer for feature extraction processing to obtain the second feature vector; The second feature vector is input into the second fully connected layer for feature extraction to obtain the third feature vector; The third feature vector is input into the normalized exponential function layer for classification processing to obtain probability values ​​corresponding to multiple preset driver intention types, and the preset driver intention type corresponding to the highest probability value is determined as the predicted driver intention type.

5. The method according to claim 1, characterized in that, The method further includes: If the vehicle is in autonomous driving mode and an abnormal signal is detected, a takeover request signal is output, and the vehicle switches to the human-machine co-driving mode. When the human-machine co-driving mode has been switched to for a preset duration, human-machine co-driving control is performed on the vehicle based on the target control parameters.

6. The method according to claim 1, characterized in that, The method further includes: If no abnormal vehicle signal is detected and an autonomous driving mode trigger signal is received, then switch to the autonomous driving mode.

7. The method according to claim 4, characterized in that, The vehicle driving data includes the vehicle's lateral speed, longitudinal speed, lateral acceleration, and longitudinal acceleration. The driver operation data includes the steering wheel angle and pedal opening values ​​operated by the driver. The vehicle dynamics model includes a two-degree-of-freedom dynamics model and a longitudinal dynamics model. Determining the vehicle driving data under driver control based on the driver operation data and the vehicle dynamics model includes: Based on the vehicle's two-degree-of-freedom dynamics model and the steering wheel angle data, the lateral travel speed is determined, and the lateral travel acceleration is determined based on the lateral travel speed. Based on the vehicle longitudinal dynamics model and the pedal opening value, the longitudinal driving speed is determined, and the longitudinal driving acceleration is determined based on the longitudinal driving speed.

8. An emergency human-machine co-driving takeover control device for an autonomous vehicle, characterized in that, The device includes: The first determining module is used to determine the predicted driver intention type based on driver operation data, vehicle external environment data, and a preset intention recognition model when the vehicle is in human-machine co-driving mode. The second determining module is used to determine the target control parameters of the vehicle based on the preset driving rights allocation strategy, vehicle operation data, and the predicted driver intention type.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images