Conflict control method and device for man-machine co-driving, storage medium and electronic equipment

Abstract

The present disclosure relates to the technical field of automatic driving, in particular to a man-machine co-driving conflict control method and device, a storage medium and an electronic device. The man-machine co-driving conflict control method comprises: establishing a man-machine path tracking control game model corresponding to man-machine interaction behavior based on a deterministic steering torque of a driver and a random steering torque of the driver; solving the man-machine path tracking control game model to obtain man-machine torque conflict information; and determining a shared control strategy according to the man-machine torque conflict information, so as to control the vehicle according to the shared control strategy. The man-machine co-driving conflict control method provided by the present disclosure can describe the mapping relationship between the man-machine decision divergence and the interaction between the man-machine steering torques, so as to improve the vehicle control accuracy.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of automatic driving, in particular to a man-machine co-driving conflict control method, a man-machine co-driving conflict control device, a storage medium and an electronic device. BACKGROUND

[0002] Since the automatic driving system and the driver are both agents, they will make judgments and decisions according to their understanding of the scene, so in addition to the man-machine steering conflict caused by the difference in man-machine preview behavior in the co-driving system control layer, another main cause of man-machine conflict is the difference in decision-making layer, that is, the difference between the target trajectory planned by the driver and the automatic driving system, which further causes the steering torque conflict.

[0003] However, in the modeling of man-machine interaction mechanism, the driver's decision-making goal is difficult to directly measure, especially in the modeling process of man-machine interaction under vehicle extreme conditions (such as man-machine cooperative emergency avoidance), the man-machine interaction behavior is difficult to directly apply a linear dynamic model to describe.

[0004] Many nonlinear methods, such as nonlinear prediction methods, local linearization methods, and piecewise affine methods, have been applied to deal with the model mismatch problem under vehicle extreme conditions. However, the nonlinear prediction method often leads to poor real-time performance of the algorithm due to its high computational complexity. Although the local linearization or piecewise affine method can ensure the real-time performance of the algorithm, these methods inevitably cause the control strategy to switch back and forth within different linearization intervals, resulting in non-smooth phenomena in the man-machine interaction results, and even the sliding mode phenomenon of switching back and forth near the linear segmentation point.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present disclosure is to provide a man-machine co-driving conflict control method, which aims to describe the mapping relationship between man-machine decision-making differences and man-machine steering torque interaction, so as to improve the control accuracy of the vehicle.

[0007] Other characteristics and advantages of the present disclosure will become apparent from the following detailed description, or will be learned by practice of the present disclosure.

[0008] According to an aspect of some embodiments of the present disclosure, a human-machine co-driving conflict control method is provided, including: establishing a human-machine path tracking control game model corresponding to human-machine interaction behaviors based on a deterministic steering torque of a driver and a random steering torque of the driver; solving the human-machine path tracking control game model to obtain human-machine torque conflict information; and determining a shared control strategy according to the human-machine torque conflict information, so as to control a vehicle according to the shared control strategy.

[0009] According to some embodiments of the present disclosure, based on the foregoing scheme, when the human-machine path tracking control game model is a closed-loop game model, establishing a human-machine path tracking control game model corresponding to human-machine interaction behaviors based on a deterministic steering torque of a driver and a random steering torque of the driver includes: establishing a first discrete state update equation of a human-machine co-driving vehicle dynamics system under a closed-loop information mode based on the deterministic steering torque of the driver and the random steering torque of the driver; augmenting the first discrete state update equation through a human-machine preview dynamic process to obtain a path tracking augmented system containing a human-machine preview state; and constructing a driver trajectory cost function and a driving system trajectory cost function based on the path tracking augmented system, so as to obtain the human-machine path tracking control game model.

[0010] According to some embodiments of the present disclosure, based on the foregoing scheme, when the human-machine path tracking control game model is an open-loop game model, establishing a human-machine path tracking control game model corresponding to human-machine interaction behaviors based on a deterministic steering torque of a driver and a random steering torque of the driver includes: establishing a second discrete state update equation of a human-machine co-driving vehicle dynamics system under an open-loop information mode based on the deterministic steering torque of the driver and the random steering torque of the driver; determining a predicted output vector within a prediction time domain according to the second discrete state update equation, and determining a driver reference trajectory vector and a driving system reference trajectory vector; and constructing a driver trajectory cost function and a driving system trajectory cost function respectively by using the predicted output vector, the driver reference trajectory vector and the driving system reference trajectory vector, so as to obtain the human-machine path tracking control game model.

[0011] According to some embodiments of the present disclosure, based on the foregoing scheme, the solving of the human-machine path tracking control game model to obtain human-machine torque conflict information includes: determining a recursive relationship of steering control value functions respectively corresponding to the driver and the driving system under a Nash equilibrium condition by using a stochastic dynamic programming algorithm; and calculating a closed-loop Nash equilibrium solution respectively corresponding to the driver and the driving system based on the first discrete state update equation and the recursive relationship as the human-machine torque conflict information.

[0012] According to some embodiments of the present disclosure, based on the foregoing scheme, the solving the human-machine path tracking control game model to obtain human-machine torque conflict information comprises: determining a recursive relationship of a steering control value function corresponding to the driver and the driving system respectively under a Stackelberg equilibrium condition by using a stochastic dynamic programming algorithm; determining a driver reaction function according to the recursive relationship of the steering control value function corresponding to the driving system; calculating an open-loop Stackelberg equilibrium solution corresponding to the driving system based on the first discrete state update equation, the driver reaction function, and the recursive relationship of the steering control value function corresponding to the driving system; and calculating an open-loop Stackelberg equilibrium solution corresponding to the driver according to the open-loop Stackelberg equilibrium solution corresponding to the driving system, as the human-machine torque conflict information.

[0013] According to some embodiments of the present disclosure, based on the foregoing scheme, the solving the human-machine path tracking control game model to obtain human-machine torque conflict information comprises: solving a model closed-form solution corresponding to the human-machine path tracking control game model to obtain a relationship expression between human-machine steering control and a target trajectory according to the model closed-form solution; and solving the relationship expression by using a convex iteration algorithm to obtain an open-loop Nash equilibrium solution corresponding to the driver and the driving system respectively as the human-machine torque conflict information.

[0014] According to some embodiments of the present disclosure, based on the foregoing scheme, the solving the human-machine path tracking control game model to obtain human-machine torque conflict information comprises: converting the driving system trajectory cost function into a driving system trajectory optimization function considering the driver reaction function; solving the driving system trajectory optimization function to obtain an open-loop Stackelberg equilibrium solution corresponding to the driving system; and calculating an open-loop Stackelberg equilibrium solution corresponding to the driver based on the open-loop Stackelberg equilibrium solution and the driver trajectory cost function, as the human-machine torque conflict information.

[0015] According to a second aspect of the embodiments of the present disclosure, a human-machine co-driving conflict control device is provided, comprising: a modeling module configured to establish a human-machine path tracking control game model corresponding to human-machine interaction behaviors based on a deterministic steering torque of a driver and a random steering torque of the driver; a solving module configured to solve the human-machine path tracking control game model to obtain human-machine torque conflict information; and an application module configured to determine a shared control strategy according to the human-machine torque conflict information, and to perform vehicle control according to the shared control strategy.

[0016] According to a third aspect of the embodiments of the present disclosure, a computer readable storage medium is provided, which stores a computer program, and the program is executed by a processor to implement the human-machine co-driving conflict control method in the above embodiments.

[0017] According to a fourth aspect of the embodiments of the present disclosure, an electronic device is provided, and has characteristics comprising: one or more processors; a storage device configured to store one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the human-machine co-driving conflict control method in the above embodiments.

[0018] The example embodiments of the present disclosure can have the following partial or all beneficial effects:

[0019] In the technical solutions provided by some embodiments of the present disclosure, the present disclosure establishes a human-machine path tracking control game model corresponding to human-machine interaction behavior based on the deterministic steering torque of the driver and the random steering torque of the driver, and solves the human-machine path tracking control game model to obtain human-machine torque conflict information for vehicle control. On the one hand, the uncertain behavior of the driver can be taken into account in human-machine path tracking control, which is more in line with the actual scene requirements and has more accurate control effect. On the other hand, the human-machine torque conflict information obtained by solving the game model is an accurate description of the human-machine interaction behavior from the decision-making disagreement to the control conflict in the human-machine co-driving model, which can overcome the human-machine decision-making confusion problem existing in the co-driving system, thereby providing a theoretical basis for co-control strategy design and further optimizing the result of vehicle control.

[0020] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0021] The drawings herein are incorporated into the specification and form a part of the specification, show embodiments consistent with the present disclosure, and together with the specification serve to explain the principles of the present disclosure. It is obvious that the drawings in the following description are only some embodiments of the present disclosure, and those skilled in the art can obtain other drawings from these drawings without creative labor. In the drawings:

[0022] Figure 1 The flowchart schematically shows a human-machine co-driving conflict control method in an example embodiment of the present disclosure;

[0023] Figure 2 The principle diagram schematically shows steering interaction of a human-machine co-driving system in an example embodiment of the present disclosure;

[0024] Figure 3 The principle diagram schematically shows a non-cooperative game of a human-machine co-driving system in an example embodiment of the present disclosure;

[0025] Figure 4 The principle diagram schematically shows a closed-loop dynamic game of a human-machine co-driving system in an example embodiment of the present disclosure;

[0026] Figure 5 A schematic diagram illustrating a principle of an open-loop dynamic game of a man-machine co-driving system in an example embodiment of the present disclosure;

[0027] Figure 6 A schematic diagram illustrating a flow of a method of constructing a closed-loop man-machine path tracking control game model in an example embodiment of the present disclosure;

[0028] Figure 7 A schematic diagram illustrating a principle of a multi-point preview mode in an example embodiment of the present disclosure;

[0029] Figure 8 A schematic diagram illustrating a flow of a method of constructing an open-loop man-machine path tracking control game model in an example embodiment of the present disclosure;

[0030] Figure 9 A schematic diagram illustrating a flow of a method of solving a closed-loop man-machine path tracking control game model in an example embodiment of the present disclosure;

[0031] Figure 10 A schematic diagram illustrating a flow of another method of solving a closed-loop man-machine path tracking control game model in an example embodiment of the present disclosure;

[0032] Figure 11 A schematic diagram illustrating a flow of another method of solving an open-loop man-machine path tracking control game model in an example embodiment of the present disclosure;

[0033] Figure 12 A schematic diagram illustrating a flow of another method of solving an open-loop man-machine path tracking control game model in an example embodiment of the present disclosure;

[0034] Figure 13 A schematic diagram illustrating a composition of a man-machine co-driving conflict control device in an example embodiment of the present disclosure;

[0035] Figure 14 A schematic diagram illustrating a computer readable storage medium in an example embodiment of the present disclosure;

[0036] Figure 15 A schematic diagram illustrating a structure of a computer system of an electronic device in an example embodiment of the present disclosure. DETAILED DESCRIPTION

[0037] Example implementations are now described with reference to the drawings. Example implementations can, however, be implemented in many different forms and should not be construed as limited to the examples set forth herein; rather, these examples are provided so that this disclosure will be thorough and complete, and will fully convey the gist of each example to those skilled in the art.

[0038] Moreover, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the

[0039] The block diagrams in the drawings show only the functionality of the features and can not imply that the functionality must be implemented in the order shown. Some of the features can be implemented in hardware, software, or firmware, or a combination thereof. Moreover, some of the features can be implemented in one or more of the entities shown in the drawings or in other entities.

[0040] The flow diagrams shown in the drawings are examples only and are not necessarily to be construed as having any prior art effect. Not all steps are required, nor are they necessarily performed in the order shown. Some steps can be combined or partially combined, and some steps can be performed in a different order than shown.

[0041] In a human-robot system, the driver and the automatic driving system will simultaneously manipulate the actuator and change the vehicle state to achieve their respective goals, and this input redundancy inevitably causes the generation of human-robot conflict interaction, which seriously affects the safety, comfort, power and fuel economy of the vehicle.

[0042] Game theory is an effective means to describe and understand the interaction and conflict between two parties in a multi-agent system, and provides an effective theoretical method for the quantitative modeling of human-robot interaction, the resolution of human-robot conflict and the inference of the real intention of the driver.

[0043] For the problem of human-robot co-driving trajectory tracking control, since the automatic driving system and the driver are both agents, they will make judgments and decisions according to their understanding of the scene. Therefore, in addition to the human-robot steering conflict caused by the difference in human-robot preview behavior at the control layer of the co-driving system, another main cause of human-robot conflict is the difference in the target trajectory planned by the driver and the automatic driving system at the decision layer, which in turn causes a steering torque conflict.

[0044] But on the one hand, in the modeling of human-machine interaction mechanism, especially in the modeling process of human-machine interaction under the extreme working condition of the vehicle (such as human-machine cooperative emergency avoidance), the human-machine interaction behavior is difficult to be directly described by a linear dynamic model, and many nonlinear methods such as nonlinear prediction method, local linearization method and piecewise affine method are applied to deal with the model mismatch problem under the extreme working condition of the vehicle, and the nonlinear prediction method often leads to poor real-time performance of the algorithm due to its high operation complexity, and the local linearization or piecewise affine method can ensure the real-time performance of the algorithm, but these methods inevitably lead to the switching of the control strategy in different linearization intervals, resulting in the non-smooth phenomenon of human-machine interaction results, and even the sliding mode phenomenon of switching near the linear segmentation point; on the other hand, for a multi-agent dynamic system, dynamic non-cooperative game theory can also be used, but the current game theory in the shared control strategy design is mainly limited to the shared control strategy design, and there is no perfect theoretical description of the mapping relationship between human-machine decision divergence and control conflict.

[0045] To solve this thorny problem, the present disclosure studies the human-machine interaction mechanism of decision divergence to control conflict. Since there are both a certain steering torque and an uncertain steering torque of the driver in the human-machine coupled steering dynamic system, the present disclosure proposes a new stochastic game theory framework, considers the deterministic and random steering torque, and approximates the Nash and Stackelberg equilibrium under different information modes to completely describe the mapping relationship, so as to design a theoretical bridge connecting human-machine decision divergence and control conflict, overcome the human-machine decision confusion problem existing in the shared driving system, and then provide a theoretical basis for the design of shared control strategy.

[0046] The implementation details of the technical solutions of the embodiments of the present disclosure are described in detail below.

[0047] Figure 1 The flowchart of the human-machine shared driving conflict control method in the exemplary embodiment of the present disclosure is schematically shown. As shown in Figure 1 The human-machine shared driving conflict control method includes steps S101 to S103:

[0048] Step S101, a human-machine path tracking control game model corresponding to human-machine interaction behavior is established based on the deterministic steering torque of the driver and the random steering torque of the driver;

[0049] Step S102, the human-machine path tracking control game model is solved to obtain human-machine torque conflict information;

[0050] Step S103, a shared control strategy is determined according to the human-machine torque conflict information, and vehicle control is performed according to the shared control strategy.

[0051] In the technical solutions provided in some embodiments of the present disclosure, the present disclosure establishes a human-machine path tracking control game model corresponding to human-machine interaction behavior based on the deterministic steering torque of the driver and the random steering torque of the driver, and solves the human-machine path tracking control game model to obtain human-machine torque conflict information for vehicle control. On the one hand, the driver's uncertain behavior can be taken into account in human-machine path tracking control, which is more in line with actual scene requirements and has more accurate control effect. On the other hand, the human-machine torque conflict information obtained by solving the game model is an accurate description of the human-machine interaction behavior from the decision-making disagreement to the control conflict in the human-machine co-driving model, which can overcome the human-machine decision-making confusion problem existing in the co-driving system, thereby providing a theoretical basis for shared control strategy design and further optimizing the result of vehicle control.

[0052] In the following, the steps of the human-machine co-driving conflict control method in the present example embodiment will be described in more detail in conjunction with the accompanying drawings and examples.

[0053] In step S101, a human-machine path tracking control game model corresponding to human-machine interaction behavior is established based on the deterministic steering torque of the driver and the random steering torque of the driver.

[0054] In one embodiment of the present disclosure, a human-machine path tracking control game model for solving human-machine interaction behavior is first constructed, which can adopt a non-cooperative game theory framework.

[0055] Figure 2 A schematic diagram illustrating the principle of steering interaction of a human-machine co-driving system in an exemplary embodiment of the present disclosure is shown. As shown in Figure 2 The driver and the automatic driving system (hereinafter referred to as the driving system) will produce corresponding steering control behavior according to their own target trajectories. At the same time, since both can perceive each other's decision through the steering system and the vehicle motion state, the steering control behavior of the driver also includes the reaction to the steering behavior of the driving system in order to achieve the target of its own decision. Therefore, when designing the controller of the driving system, the steering control input of the driver should also be fully considered. Therefore, in the human-machine co-driving system, the modeling of human-machine steering conflict must consider the interaction of human-machine steering behavior, especially under the condition of disagreement between human-machine decision-making targets, the steering behavior of the driver will not only be based on its own decision-making target, but also actively compensate for the steering behavior of the automatic driving system, which is quite different from pure manual driving.

[0056] Since the steering behaviors of the driver and the automatic driving system in the co-driving process are both aimed at minimizing the trajectory tracking error, and each produces its own steering action according to the feedback of the vehicle-road system state and the steering action of the other party, the target trajectories of the human-machine decision-making layers inevitably differ. The human-machine interaction process under such conditions can be described by a non-cooperative game theory framework.

[0057] Figure 3 This illustration schematically depicts the principle of a non-cooperative game in a human-machine co-driving system according to an exemplary embodiment of this disclosure. Figure 3 As shown, the driver and the driving system are regarded as game participants, and both parties aim to maximize their own interests. They generate their own actions by combining the state information of the game process and the other party's actions.

[0058] When studying human-machine co-driving problems using game theory, the trajectory tracking behavior of the driver and the autonomous driving system is described using optimal control strategies based on cost functions. Since both cost functions are always non-zero, and optimal trajectory tracking control is typically described using prediction and control time domains, the trajectory tracking problem in human-machine co-driving can be abstracted as a non-zero-sum, multi-stage dynamic game problem. For non-zero-sum multi-stage dynamic games, the game problem can be divided into two categories based on the information pattern of the game: closed-loop memoryless dynamic game problems and open-loop dynamic game problems.

[0059] Figure 4 This diagram schematically illustrates a closed-loop dynamic game principle of a human-machine co-driving system according to an exemplary embodiment of this disclosure. Figure 4 As shown, in this information model, the permissible policy set of the participants is mapped through the initial state and the states of each stage. "Memoryless" means that the participant only knows the initial and current states of the system when making a decision at each stage, and has no memory of the states of the other stages. Therefore, the participant's action in stage i can be expressed as... i∈{0,1,…,n} u}

[0060] The counterpart to closed-loop memoryless dynamic games is open-loop dynamic games. Figure 5 This illustration schematically depicts the principle of an open-loop dynamic game in a human-machine co-driving system according to an exemplary embodiment of this disclosure. Figure 5 As shown, the entire game process is divided into n parts. u There are several stages, and the state vector of the system in each stage is x. k The permissible policy set of each participant in each stage depends only on the initial state. Therefore, when the initial state x0 is given, the policy set is a constant function, and the participant's action in stage i also becomes a constant. i∈{0,1,…,n} u}

[0061] Therefore, in step S101, two different information modes, namely closed-loop and open-loop, can be used to construct human-machine path tracking control game models.

[0062] ① Closed-loop human-machine path tracking control game model

[0063] Figure 6 Fig. 1 schematically shows a flowchart of a method for constructing a closed-loop human-machine path tracking control game model in an exemplary embodiment of the present disclosure, i.e., a linear quadratic regulator (LQR) method with optimal multi-point preview is used to describe the steering behavior of the driver and the driving system. As shown in Fig. 1, the method for constructing the human-machine path tracking control game model includes the following steps: Figure 6

[0064] Step S601, establishing a first discrete state update equation of a human-machine co-driving vehicle dynamics system in a closed-loop information mode based on the deterministic steering torque of the driver and the random steering torque of the driver;

[0065] Step S602, augmenting the first discrete state update equation through a human-machine preview dynamic process to obtain a path tracking augmented system containing a human-machine preview state;

[0066] Step S603, constructing a driver trajectory cost function and a driving system trajectory cost function based on the path tracking augmented system to obtain the human-machine path tracking control game model.

[0067] Specifically, this section focuses on the human-machine interaction modeling problem, so it can be assumed that the human-machine target trajectories have small tangent direction angles, and under the condition of small heading angle, the state vector in the model can be simplified as where θ sw is the global longitudinal coordinate of the vehicle mass center, ψ is the vehicle heading angle, sw is the derivative of ψ with respect to time. is the derivative of ψ with respect to time.

[0068] In step S601, a first discrete state update equation of a human-machine co-driving vehicle dynamics system in a closed-loop information mode is established based on the deterministic steering torque of the driver and the random steering torque of the driver.

[0069] The deterministic steering torque of the driver is denoted as τ h , the random steering torque of the driver is denoted as τ hsto , and the steering torque of the driving system is denoted as τ m . Based on the state vector x c , the deterministic steering torque of the driver τ h , and the random steering torque of the driver τ hsto , a continuous state space equation is established as shown in formula (1):

[0070]

[0071] where τ disk ​​For the distance-related steering torque, A c B h B m N c And C c These are the parameter matrices in the model, as shown below:

[0072]

[0073] To describe the human-machine co-driving problem as a multi-stage game, we use the system's discrete time T... s Discretizing the above continuous system, the human-machine shared vehicle dynamics system can be transformed into the following difference equation, namely the first discrete state update equation, as shown in equation (2):

[0074]

[0075] In the formula: For the deterministic steering torque of the driver at time k in a human-machine co-driving system, Let τ be the steering torque of the driving system at time k. hsto τ represents the driver's random steering torque. disk This refers to the steering torque, which is related to distance.

[0076] In step S602, the first discrete state update equation is augmented through the human-machine pre-aiming dynamic process to obtain a path tracking augmentation system that includes the human-machine pre-aiming state.

[0077] Figure 7 This illustration schematically demonstrates the principle of a multi-point anti-aiming mode in an exemplary embodiment of this disclosure. To model the path tracking control system using the LQR-based method and consider the human-machine predictive behavior of vehicle dynamics, the human-machine anti-aiming behavior is first modeled as follows: Figure 7 The multi-point aiming mode shown. (Reference) Figure 7 As shown, at each moment, the driver and the driving system pre-aim at a region of the target trajectory based on their own decisions. This region can be described as n p There are several aiming points, and their aiming distance is determined by the driver's aiming time t. p Decision, and t p =n p ×T s This dynamic process can be expressed using a shift register, as shown in equation (3):

[0078]

[0079] In the formula: r x(k+i) =[Y x (k+i)ψ x (k+i)]T , x∈{h,m},

[0080]

[0081] Through the man-machine preview dynamic process Augmenting the human-machine shared vehicle dynamics system can obtain a path tracking augmented system containing the man-machine preview state, as shown in equation (4):

[0082]

[0083] In the formula (4): x k =[x ck R hk R mk ] T ,

[0084]

[0085] In equation (4), let c k =Nτ disk , The lateral displacement and heading angle of the farthest preview point of the preview area of the driver and the driving system, since the preview information of the driver and the driving system in the remaining area is located in the augmented state x k , the farthest preview point information The path tracking augmented system can be further simplified as shown in equation (5):

[0086]

[0087] Step S603, based on the path tracking augmented system, construct a driver trajectory cost function and a driving system trajectory cost function to obtain the human-machine path tracking control game model.

[0088] Specifically, in the path tracking augmented system where the human-machine decision diverges, the driver trajectory cost function J1 and the driving system trajectory cost function J2 are designed with a prediction and control time domain of n u Step length to obtain the human-machine path tracking control game model can be described as shown in equation (6):

[0089]

[0090] In the formula:

[0091]

[0092]

[0093] Where Q1 and Q2 are the state weighting matrices of the driver and the driving system, respectively, and W1 and W2 are the tracking error weighting matrices of the driver and the driving system, respectively. These are the weighting coefficients for the driver's vehicle lateral tracking error and heading angle error, respectively. R represents the weighting coefficients for the vehicle lateral tracking error and heading angle error of the driving system, respectively. 11 R 22 R represents the input weighting coefficients for the driver and the driving system, respectively. 11 =q u1 R 22 =q u2 R 12 R 21 R represents the weighting coefficients for the interaction inputs between the driver and the driving system. 12 =q u12 R 21 =q u21 .

[0094] Based on this, formula (6) establishes n using the linear quadratic form method. u In this phase of the human-machine path tracking control game model, the cost functions of both sides include the other party's turning control input to express the human-machine interaction characteristics.

[0095] ② Open-loop human-machine path tracking control game model

[0096] Figure 8 This illustration schematically depicts a flowchart of a method for constructing an open-loop human-machine path tracking control game model according to an exemplary embodiment of this disclosure. Specifically, it employs a Distributed Model Predictive Control (DMPC) strategy to describe the mapping relationship between human-machine decision disagreements and steering torque conflicts under an open-loop information model. Figure 8 As shown, the method for constructing a human-machine path tracking control game model includes the following steps:

[0097] Step S801: Based on the driver's deterministic steering torque and the driver's stochastic steering torque, establish the second discrete state update equation of the human-machine co-driving vehicle dynamics system in open-loop information mode;

[0098] Step S802: Determine the prediction output vector in the prediction time domain according to the second discrete state update equation, and determine the driver reference trajectory vector and the driving system reference trajectory vector;

[0099] Step S803, using the predicted output vector, the driver reference trajectory vector and the driving system reference trajectory vector to construct a driver trajectory cost function and a driving system trajectory cost function respectively to obtain the human-machine path tracking control game model.

[0100] Specifically, under the model predictive control framework, both the driver and the driving system estimate the vehicle operation trajectory within the prediction horizon n p , and apply steering control within the control horizon n u to minimize the deviation between the vehicle trajectory and the respective decision, compared with the linear quadratic regulator method, the model predictive control more intuitively reflects the target trajectory of the driver and the automatic driving system decision layer planning in the final interaction model. At the same time, according to the model predictive control algorithm, the establishment of the state prediction vector in the cost function is based on the current state of the system and the control input within the control horizon n u , so the control law is only related to the current initial state, which exactly meets the definition of open-loop information mode.

[0101] In step S801, a second discrete state update equation of the human-machine co-driving vehicle dynamics system in the open-loop information mode is established based on the deterministic steering torque of the driver and the random steering torque of the driver.

[0102] For the human-machine co-driving vehicle dynamics system, the driver's uncertain behavior is incorporated into the system disturbance, and the second discrete state update equation is as shown in formula (7):

[0103]

[0104] In the formula: is the deterministic steering torque of the driver of the human-machine co-driving system at time k, is the steering torque of the driving system at time k, τ hsto is the random steering torque of the driver, τ disk is the distance-dependent steering torque, and the disturbance input w’ k is recorded. disk hsto .

[0105] Based on the above second discrete state update equation, if it is assumed that the disturbance input w’ k remains unchanged within the prediction horizon, then the model output of the human-machine co-driving system in the next n p step can be expressed as:

[0106]

[0107] ​Step S802, determining a predicted output vector in a prediction horizon according to the second discrete state update equation, and determining a driver reference trajectory vector and a driving system reference trajectory vector.

[0108] It is assumed that the prediction horizon and the control horizon of the model prediction algorithm of the man-machine co-driving system are both n u At time k, define the model prediction output vector in the prediction horizon as Y pk As shown in equation (8), the man-machine control input vectors are U hk and U mk respectively as shown in equation (9) and equation (10):

[0109]

[0110]

[0111]

[0112] Based on the second discrete state update equation (equation 7), the model output of the man-machine co-driving system in the next n p step, i.e. the predicted output vector, can be expressed as equation (11):

[0113] Y pk = Ψx ck + Θ h U hk + Θ m U mk + Sw’ k (11)

[0114] In the equation:

[0115]

[0116]

[0117]

[0118] At each time step k, the reference trajectory vectors of the driver and the automatic driving system can be expressed as equation (12):

[0119]

[0120] Step S803, constructing a driver trajectory cost function and a driving system trajectory cost function respectively using the predicted output vector, the driver reference trajectory vector and the driving system reference trajectory vector to obtain the man-machine path tracking control game model.

[0121] The parameters in the human-machine path tracking control game model are predicted output vectors Y pk and reference trajectory vectors R hk , R mk , where the driver trajectory cost function and the driving system trajectory cost function are obtained as shown in equation (13):

[0122]

[0123] where Γ yx , Γ ux are weighting matrices of path tracking control, and x e {1, 2},

[0124]

[0125] In step S102, human-machine torque conflict information is obtained by solving the human-machine path tracking control game model.

[0126] Since the human-machine path tracking control game models in the closed loop and the open loop are constructed in step S101, the solving methods for the two game models are different.

[0127] In addition, for non-cooperative game problems, since the human-machine co-driving system usually adopts a mirror-symmetrical personification strategy to design an automatic driving system controller to improve human-machine consistency, the driver and the automatic driving system have an equal relationship in the game process. When the game participants are in a symmetric or equal relationship, Nash Equilibrium provides a reasonable theoretical solution for non-cooperative games. In this equilibrium state, neither party dominates the decision-making process, and neither party can reduce its trajectory tracking cost function value by adjusting its own decision-making.

[0128] In addition, there is another type of non-cooperative game problem, i.e., the master-slave game problem of a double-agent system, in which the leader knows the reaction of the follower to his decision and makes the decision first, and the follower makes the corresponding decision after observing the decision of the leader. This hierarchical game result is called Stackelberg Equilibrium.

[0129] Since Nash equilibrium and Stackelberg equilibrium can model the decision-control mechanism of the human-machine co-driving system theoretically, for each game model, the Nash equilibrium and Stackelberg equilibrium corresponding to the model can be solved. Therefore, in the following, four kinds of quantitative models based on non-cooperative game theory are adopted, based on open-loop and closed-loop information modes, to deeply analyze the mapping relationship between the human-machine decision difference and the steering torque conflict under the Nash equilibrium and Stackelberg equilibrium, so as to explore the best modeling method of the human-machine interaction mechanism.

[0130] ① Nash equilibrium solution of human-machine decision-control under closed-loop information mode

[0131] In an embodiment of the present disclosure, the solution problem of Nash equilibrium under closed-loop information mode can be described as a Hamilton function constraint optimization problem as shown in formula (14):

[0132]

[0133] In the formula: is the coordination vector.

[0134] Closed-form solution of the constraint optimization problem constitutes the closed-loop Nash equilibrium solution.

[0135] The Nash equilibrium solution under the open-loop mode is only related to the initial state and is irrelevant to the current state of each stage, so the open-loop Nash equilibrium solution also satisfies the closed-loop Nash equilibrium condition (14), but obviously the solution of the closed-loop Nash equilibrium problem is not only the open-loop Nash equilibrium solution, and this information non-uniqueness leads to the multi-solution problem of Nash equilibrium under the closed-loop information structure. Since the parameter θ k of the driver's random steering torque is considered in the system modeling in this paper, the information non-uniqueness in the game process is eliminated, so the feedback Nash equilibrium can be used to avoid the multi-solution problem, and the optimal solution of each stage of the feedback Nash equilibrium satisfies the following conditions:

[0136]

[0137] It can be seen that the feedback Nash equilibrium condition under the closed-loop condition conforms to the Bellman optimality principle, and the closed-loop Nash equilibrium solution (also called feedback Nash equilibrium solution) of the path tracking control system under the human-machine decision difference condition can be solved by the Stochastic Dynamic Programming (SDP) algorithm.

[0138] Figure 9 A flowchart schematically showing a method for solving a closed-loop human-machine path tracking control game model in an exemplary embodiment of the present disclosure is shown in FIG. 6. As shown in FIG. 6, the method comprises the following steps: Figure 9As shown, the solving human-machine path tracking control game model method includes the following steps:

[0139] Step S901, a random dynamic programming algorithm is used to determine the recursive relationship of the steering control value function corresponding to the driver and the driving system respectively under the Nash equilibrium condition;

[0140] Step S902, based on the first discrete state update equation and the recursive relationship, the closed-loop Nash equilibrium solution corresponding to the driver and the driving system respectively is calculated as the human-machine torque conflict information.

[0141] Specifically, at any time step k, it is assumed that the human-machine co-driving system has unmodeled disturbance c k In the entire n u The phase dynamic game remains unchanged, and the Gaussian random distribution θ k ~N(μ,σ) is used to describe the parameters about the random steering torque of the driver, wherein μ and σ represent the mean and standard deviation of the Gaussian distribution respectively.

[0142] Due to the existence of the disturbance vectors c k and θ k , the solution of the human-machine game problem based on the steering torque interaction becomes an affine quadratic form, so in the dynamic programming solving process, the steering control value function of the driver and the driving system has the following affine quadratic form as shown in formula (15):

[0143]

[0144] The steering control value function in the above formula (15) represents the cost function of the driver and the driving system respectively, which is the value of the value function in the game process from the jth stage to the nth u stage, and the steering control value functions of the k+j step and the k+j+1 step satisfy the recursive relationship as shown in formula (16):

[0145]

[0146] In the formula, the superscript "N" represents the value function under the Nash equilibrium condition, and the single-step cost function g() of each stage can be expressed as formula (17):

[0147]

[0148] Substitute the first discrete state update equation (Formula 4) into the steering control value function (Formula 15) to obtain the human-machine torque relationship satisfying the closed-loop Nash equilibrium relationship as shown in formula (18):

[0149]

[0150] 2. Closed-loop information mode human-machine decision-making-control Stackelberg equilibrium solution

[0151] Figure 10 Fig. 2 schematically shows a flowchart of another method for solving a closed-loop human-machine path tracking control game model in an exemplary embodiment of the present disclosure. As shown in Fig. 2, the method for solving a human-machine path tracking control game model includes the following steps: Figure 10

[0152] Step S1001, using a stochastic dynamic programming algorithm to determine the recursive relationship of the steering control value function corresponding to the driver and the driving system under the Stackelberg equilibrium condition;

[0153] Step S1002, determining the driver reaction function according to the recursive relationship of the steering control value function corresponding to the driving system;

[0154] Step S1003, calculating the open-loop Stackelberg equilibrium solution corresponding to the driving system based on the first discrete state update equation, the driver reaction function, and the recursive relationship of the steering control value function corresponding to the driving system;

[0155] Step S1004, calculating the open-loop Stackelberg equilibrium solution corresponding to the driver according to the open-loop Stackelberg equilibrium solution corresponding to the driving system, as the human-machine torque conflict information.

[0156] Specifically, unlike Nash equilibrium, in Stackelberg equilibrium, there is a master-slave relationship between the driver and the automatic driving system. In each stage of the game, the driving system as the leader first takes the steering control input, and the driver observes this action and makes a corresponding response.

[0157] Therefore, the closed-loop Stackelberg equilibrium solution can be solved by backward induction. Based on the stochastic dynamic programming method, the recursive relationship of the driver steering control value function is the same as (see equation 16).

[0158] And the recursive relationship of the steering control value function of the driving system in the closed-loop Stackelberg equilibrium solution (also known as the feedback Stackelberg equilibrium solution) under the closed-loop information mode satisfies the condition as shown in equation (19):

[0159]

[0160] In the formula, the superscript "S" represents the value function under the Stackelberg equilibrium condition, wherein

[0161]

[0162] ​The first discrete state update equation The driver steering control value function recursive relation Substituting the driver steering control value function recursive relation (19) with the first discrete state update equation (4) and the driver reaction function (20), the open-loop Stackelberg equilibrium solution of the driver-vehicle system under the Stackelberg equilibrium strategy is shown in equation (21):

[0163]

[0164] Substituting the driver steering control value function recursive relation (19) with the first discrete state update equation (4) and the driver reaction function (20), the open-loop Stackelberg equilibrium solution of the driver-vehicle system under the Stackelberg equilibrium strategy is shown in equation (21):

[0165]

[0166] In the equation,

[0167] Since the driver-vehicle system is the leader and the driver is the follower, the open-loop Stackelberg equilibrium solution of the driver-vehicle system can be subsequently used to solve the corresponding Hamilton function constraint optimization problem of the driver to obtain a closed-form solution as the open-loop Stackelberg equilibrium solution of the driver.

[0168] ③ Open-loop information mode human-machine decision-control Nash equilibrium solution

[0169] Figure 11 An exemplary flowchart schematically showing another method for solving the open-loop human-machine path tracking control game model in the exemplary embodiments of the present disclosure is shown in FIG. 11. As shown in FIG. 11, the method for solving the human-machine path tracking control game model includes the following steps: Figure 11

[0170] Step S1101, solving the model closed-form solution corresponding to the human-machine path tracking control game model to obtain a relationship expression between the human-machine steering control and the target trajectory according to the model closed-form solution;

[0171] Step S1102, solving the relationship expression using a convex iteration algorithm to obtain the open-loop Nash equilibrium solution corresponding to the driver and the driver-vehicle system respectively as the human-machine torque conflict information.

[0172] Specifically, based on the definition of the open-loop Nash equilibrium, the convex iteration algorithm is used to solve the Nash equilibrium solution of the system under the open-loop information mode.

[0173] For the unconstrained problem as shown in equation (13), the human-machine path tracking control law has the following closed-form solution, as shown in equation (22):

[0174] ​

[0175] wherein:

[0176]

[0177] wherein, pinv() is pseudo-inverse operator, pinv(A) = (A T A) -1 A T Thus, the relationship between human-machine steering control and target trajectory is obtained, see equation (23):

[0178]

[0179] From equation (23), it can be seen that the control law of the two exists interactive coupling relationship, that is, the control behavior is not only related to the decision target, system state, disturbance, but also closely related to the control strategy of the other party. Therefore, in order to realize the decoupling of equation (23), the convex iteration algorithm is used to solve it, and the update formula is obtained as shown in equation (24):

[0180]

[0181] The solving process of convex iteration algorithm can be summarized as follows: first, determine the initial value of iteration U hk(0) , U mk(0) , then substitute it into the relationship expression (equation 23) to obtain the current optimal value Substitute it into the update formula (equation 24) to update the next step U hk(1) , U mk(1) , and so on. When i tends to infinity, the update formula (equation 24) becomes:

[0182]

[0183] Therefore, the relationship expression (equation 23) becomes:

[0184]

[0185] Therefore, the expression of human-machine decision-control Nash equilibrium solution in open-loop information mode is shown in equation (27):

[0186]

[0187] In the actual modeling process, the value of the first stage of the game is taken as the open-loop Nash equilibrium of human-machine interaction at time k, that is, the human-machine torque conflict information:

[0188]

[0189] ④Open-loop information mode of human-machine decision-control Stackelberg equilibrium solution

[0190] Figure 12 Fig. 6 schematically shows a flowchart of another method for solving the open-loop human-machine path tracking control game model according to an example embodiment of the present disclosure. As shown in Fig. 6, the method for solving the human-machine path tracking control game model comprises the following steps: Figure 12

[0191] Step S1201, converting the driving system trajectory cost function into a driving system trajectory optimization function considering the driver reaction function;

[0192] Step S1202, solving the driving system trajectory optimization function to obtain an open-loop Stackelberg equilibrium solution corresponding to the driving system;

[0193] Step S1203, calculating an open-loop Stackelberg equilibrium solution corresponding to the driver based on the open-loop Stackelberg equilibrium solution and the driver trajectory cost function, and taking the open-loop Stackelberg equilibrium solution as the human-machine torque conflict information.

[0194] Specifically, the open-loop Stackelberg equilibrium solution is realized by using a reverse induction method similar to the closed-loop Stackelberg equilibrium solution. The predicted output vector Y pk Substituting equation (11) into the game model (13) gives:

[0195]

[0196] In the equation, E hk = R hk - Ψx ck - S d w k - Θ m U mk , E mk = R mk - Ψx ck - S d w k - Θ h U hk .

[0197] Unlike the open-loop Nash equilibrium solution, in the master-slave game, the autonomous driving system acts as the leader and takes the first action in each stage of the game process, and the driver makes a corresponding response after observing the action of the autonomous driving system.

[0198] Therefore, the reaction function of the driver to the driving system is obtained by solving J1, as shown in equation (30):

[0199]

[0200] ​When the control strategy is taken, the driving system will consider the driver's reaction function to it into its cost function to get a new driving system trajectory cost function, as shown in equation (31):

[0201] J2=||Γ y2 [Θ m U mk -R mk +Ψx ck +S d w’ k +Θ h L hk E hk ]|| 2 +||Γ u2 U mk || 2 (31)

[0202] Therefore, solving the open-loop information mode driving system steering torque control Stackelberg equilibrium solution becomes to solve the unconstrained optimization problem as shown in equation (32):

[0203]

[0204] By solving the above optimization problem, the corresponding open-loop Stackelberg equilibrium solution of the driving system can be obtained as shown in equation (33):

[0205]

[0206] In the formula: Ψ' m =Ψ-Θ h L hk Ψ,S' m =S-Θ h L h S d ,Θ' m =Θ h L h ,Θ' m1 =Θ m -Θ h L h Θ m ,

[0207] Stackelberg equilibrium solution based on driving system steering control The optimal response of the driver to the automatic driving system steering control can be obtained by solving the optimization problem as shown in equation (34):

[0208]

[0209] The open-loop information mode driver steering control Stackelberg equilibrium solution can be obtained from the above formula, as shown in formula (35):

[0210]

[0211] In the formula, T1'=I+Θ m L' m Θ' m , Θ' h =Θ m L' m , Ψ'1=-Ψ+Θ m L' m Ψ' m , S'1=-S+Θ m L' m S' m

[0212] Similarly, the open-loop Stackelberg equilibrium solution of the first stage of the game describes the human-machine steering torque interaction at time k as shown in formula (36):

[0213]

[0214] Unlike the open-loop Nash equilibrium, because the open-loop Stackelberg equilibrium solution is based on the backward induction method, when solving the automatic driving system control law (formula 34), it can also be changed into a constrained optimization problem by adding constraints, so that the controller meets the expected performance. Therefore, the significance of the open-loop Stackelberg equilibrium strategy proposed in this paper is not only to propose a theoretical model of human-machine steering torque interaction, but also to apply the algorithm to flexibly design an interactive steering assistance controller that meets kinematic and dynamic safety constraints.

[0215] In step S103, a shared control strategy is determined according to the human-machine torque conflict information, so as to control the vehicle according to the shared control strategy.

[0216] Specifically, the human-machine torque conflict information obtained in step S102 establishes a theoretical basis between human-machine decision divergence and control conflict, based on which the shared control strategy design under the emergency lane changing condition can be better designed to guide the vehicle control.

[0217] The present disclosure establishes a theoretical relationship between human-machine decision divergence and control conflict, and models the mapping relationship between human-machine decision divergence and human-machine steering torque interaction through four kinds of dynamic non-cooperative game theoretical frameworks.

[0218] In one aspect, the human-machine interaction behavior in the feedback information mode is modeled by augmenting the human-machine shared control system dynamics model with the human-machine target trajectory, and then by using the linear quadratic regulator method.

[0219] Since the model establishes the mapping relationship between the human-machine decision divergence and the steering torque interaction, the existence of the steering resistance torque and the driver's uncertain torque makes the solution of the human-machine decision divergence problem become an affine quadratic game problem, therefore, this paper proposes an affine quadratic game algorithm based on stochastic dynamic programming, and solves the feedback Nash equilibrium solution and the feedback Stackelberg equilibrium solution describing the interaction relationship between human-machine decision and steering torque.

[0220] On the other hand, the distributed model predictive control method is used to describe the human-vehicle multi-objective path tracking control problem in the open-loop information mode, in order to solve the human-machine decision-control model in the open-loop information mode, the open-loop Nash equilibrium solution and the open-loop Stackelberg equilibrium solution describing the mapping relationship between human-machine decision and control are further solved by the model predictive control method.

[0221] Figure 13 The composition of a human-machine shared control conflict device in an exemplary embodiment of the present disclosure is schematically shown as follows: Figure 13 As shown in the figure, the human-machine shared control conflict device 1300 can include a modeling module 1301, a solving module 1302, and an application module 1303. Among them:

[0222] The modeling module 1301 is configured to establish a human-machine path tracking control game model corresponding to human-machine interaction behavior based on the driver's deterministic steering torque and the driver's random steering torque;

[0223] The solving module 1302 is configured to solve the human-machine path tracking control game model to obtain human-machine torque conflict information;

[0224] The application module 1303 is configured to determine a shared control strategy according to the human-machine torque conflict information, and to control the vehicle according to the shared control strategy.

[0225] According to the exemplary embodiment of the present disclosure, the modeling module 1301 includes a first modeling unit configured to establish a first discrete state update equation of a human-machine shared control vehicle dynamics system in a closed-loop information mode based on the driver's deterministic steering torque and the driver's random steering torque; augment the first discrete state update equation through a human-machine preview dynamic process to obtain a path tracking augmented system containing human-machine preview state; and construct a driver trajectory cost function and a driving system trajectory cost function based on the path tracking augmented system to obtain the human-machine path tracking control game model.

[0226] According to an example embodiment of the present disclosure, the modeling module 1301 comprises a second modeling unit configured to establish a second discrete state update equation of a human-machine co-driving vehicle dynamics system in an open-loop information mode based on the deterministic steering torque of the driver and the random steering torque of the driver; determine a predicted output vector in a prediction time domain according to the second discrete state update equation, and determine a driver reference trajectory vector and a driving system reference trajectory vector; and construct a driver trajectory cost function and a driving system trajectory cost function respectively by using the predicted output vector, the driver reference trajectory vector and the driving system reference trajectory vector, so as to obtain the human-machine path tracking control game model.

[0227] According to an example embodiment of the present disclosure, the solving module 1302 comprises a first solving unit configured to determine a recursive relationship of the steering control value function corresponding to the driver and the driving system respectively under a Nash equilibrium condition by using a stochastic dynamic programming algorithm; and calculate a closed-loop Nash equilibrium solution corresponding to the driver and the driving system respectively based on the first discrete state update equation and the recursive relationship, as the human-machine torque conflict information.

[0228] According to an example embodiment of the present disclosure, the solving module 1302 comprises a second solving unit configured to determine a recursive relationship of the steering control value function corresponding to the driver and the driving system respectively under a Stackelberg equilibrium condition by using a stochastic dynamic programming algorithm; determine a driver reaction function according to the recursive relationship of the steering control value function corresponding to the driving system; calculate an open-loop Stackelberg equilibrium solution corresponding to the driving system based on the first discrete state update equation, the driver reaction function and the recursive relationship of the steering control value function corresponding to the driving system; and calculate an open-loop Stackelberg equilibrium solution corresponding to the driver according to the open-loop Stackelberg equilibrium solution corresponding to the driving system, as the human-machine torque conflict information.

[0229] According to an example embodiment of the present disclosure, the solving module 1302 comprises a third solving unit configured to solve a model closed-form solution corresponding to the human-machine path tracking control game model, so as to obtain a relationship expression between the human-machine steering control and the target trajectory according to the model closed-form solution; and solve the relationship expression by using a convex iteration algorithm to obtain an open-loop Nash equilibrium solution corresponding to the driver and the driving system respectively, as the human-machine torque conflict information.

[0230] According to an example embodiment of the present disclosure, the solving module 1302 comprises a fourth solving unit configured to convert the driving system trajectory cost function into a driving system trajectory optimization function considering the driver reaction function; solve the driving system trajectory optimization function to obtain an open-loop Stackelberg equilibrium solution corresponding to the driving system; and calculate an open-loop Stackelberg equilibrium solution corresponding to the driver based on the open-loop Stackelberg equilibrium solution and the driver trajectory cost function, as the human-machine torque conflict information.

[0231] The specific details of each module in the man-machine co-pilot conflict control device 1300 described above have been described in detail in the corresponding man-machine co-pilot conflict control method, and therefore will not be described here.

[0232] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, such division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into several modules or units embodied.

[0233] In the exemplary embodiments of the present disclosure, a storage medium capable of implementing the above method is also provided. Figure 14 The schematic diagram of a computer readable storage medium in the exemplary embodiments of the present disclosure is schematically shown as Figure 14 As shown, a program product 1400 for implementing the above method according to the embodiments of the present disclosure is described, which can adopt a portable compact disc read-only memory (CD-ROM) and include program codes, and can run on a terminal device such as a mobile phone. However, the program product of the present disclosure is not limited thereto, and in this document, the readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or apparatus.

[0234] In the exemplary embodiments of the present disclosure, an electronic device capable of implementing the above method is also provided. Figure 15 The structural schematic diagram of a computer system of an electronic device in the exemplary embodiments of the present disclosure is schematically shown.

[0235] It should be noted that, Figure 15 The computer system 1500 of the electronic device shown is only an example, and should not bring any limitation to the functions and use range of the embodiments of the present disclosure.

[0236] As Figure 15As shown, the computer system 1500 includes a central processing unit (CPU) 1501 which can perform various suitable actions and processes in accordance with programs stored in a read-only memory (ROM) 1502 or loaded from the storage section 1508 into a random access memory (RAM) 1503. Various programs and data required for system operation are also stored in the RAM 1503. The CPU 1501, the ROM 1502, and the RAM 1503 are connected to each other through a bus 1504. An input / output (I / O) interface 1505 is also connected to the bus 1504.

[0237] Connected to the I / O interface 1505 are an input section 1506 including a keyboard, a mouse, etc.; an output section 1507 including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1508 including a hard disk, etc.; and a communication section 1509 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 1509 performs communication processing via a network such as the Internet. A drive 1510 is also connected to the I / O interface 1505 as necessary. A removable recording medium 1511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is attached to the drive 1510 as necessary, so that a computer program read therefrom is installed in the storage section 1508 as necessary.

[0238] In particular, according to embodiments of the present disclosure, the processes described below with reference to the flowcharts can be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer readable medium, the computer program containing program code for executing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network by the communication section 1509, and / or installed from the removable recording medium 1511. When the computer program is executed by the central processing unit (CPU) 1501, various functions defined in the system of the present disclosure are performed.

[0239] It should be noted that the computer-readable medium in the present disclosure can be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. The computer-readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or apparatus, or any combination thereof. More specific examples of the computer-readable storage medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (Compact Disc Read-Only Memory, CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or apparatus. In the present disclosure, the computer-readable signal medium can include a data signal carrying a computer-readable program code in a baseband or as a part of a carrier wave. Such a propagated data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, which can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or apparatus. The program code contained in the computer-readable medium can be transmitted by any suitable medium, including but not limited to wireless, wired, or the like, or any suitable combination thereof.

[0240] The flowcharts and block diagrams in the drawings illustrate the possible implementation architectures, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowcharts or block diagrams can represent a module, a program segment, or a portion of code that contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur in different orders than that shown in the drawings. For example, two blocks that are shown in succession can actually be executed substantially concurrently, or they can sometimes be executed in reverse order, depending on the functionality involved. It should also be noted that each block in the block diagrams or flowcharts, and combinations of blocks in the block diagrams or flowcharts, can be implemented by a dedicated hardware-based system that performs specified functions or operations, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0241] The units described in the embodiments of the present disclosure can be implemented in the form of software, or can be implemented in the form of hardware, and the described units can also be arranged in a processor. In some cases, the names of these units do not constitute a limitation on the units themselves.

[0242] As another aspect, the present disclosure also provides a computer readable medium, which can be included in the electronic device described in the above embodiments, or can exist independently without being assembled into the electronic device. The above computer readable medium carries one or more programs, which, when executed by the electronic device, enable the electronic device to implement the method described in the above embodiments.

[0243] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, such division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into a plurality of modules or units.

[0244] From the above description of the embodiments, those skilled in the art can easily understand that the example embodiments described herein can be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product. The software product can be stored in a non-volatile storage medium (which can be a CD-ROM, a U disk, a mobile hard disk, etc.) or a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the methods according to the embodiments of the present disclosure.

[0245] Other embodiments of the present disclosure will be apparent to those skilled in the art with the disclosure herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure following, in general, the principles of the present disclosure and including such departures from the present disclosure that come within known or customary practice within the art to which the present disclosure pertains.

[0246] It should be understood that the present disclosure is not limited to the precise structures described and shown in the above and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for controlling conflicts in human-machine co-driving, characterized in that, The method comprises the following steps: establishing a human-machine path tracking control game model corresponding to human-machine interaction behaviors based on a deterministic steering torque of a driver and a random steering torque of the driver; the human-machine path tracking control game model comprises a driver trajectory cost function and a driving system trajectory cost function; solving the human-machine path tracking control game model to obtain human-machine torque conflict information; determining a shared control strategy according to the human-machine torque conflict information, and performing vehicle control according to the shared control strategy; wherein, when the human-machine path tracking control game model is a closed-loop game model, the step of establishing the human-machine path tracking control game model corresponding to human-machine interaction behaviors based on the deterministic steering torque of the driver and the random steering torque of the driver comprises: establishing a first discrete state update equation of a human-machine co-driving vehicle dynamics system in a closed-loop information mode based on the deterministic steering torque of the driver and the random steering torque of the driver; augmenting the first discrete state update equation through a human-machine preview dynamic process to obtain a path tracking augmented system containing a human-machine preview state; constructing the driver trajectory cost function and the driving system trajectory cost function based on the path tracking augmented system to obtain the human-machine path tracking control game model; when the human-machine path tracking control game model is an open-loop game model, the driver trajectory cost function and the driving system trajectory cost function are as follows: wherein: , are weighting matrices of the path following control, is a model predicted output vector in a prediction horizon, , are reference trajectory vectors of the driver and the autonomous driving system, respectively, , are control input vectors of the driver and the autonomous driving system, respectively.

2. The method according to claim 1, wherein, when the human-machine path tracking control game model is an open-loop game model, the step of establishing the human-machine path tracking control game model corresponding to human-machine interaction behaviors based on the deterministic steering torque of the driver and the random steering torque of the driver comprises: establishing a second discrete state update equation of the human-machine co-driving vehicle dynamics system in an open-loop information mode based on the deterministic steering torque of the driver and the random steering torque of the driver; determining a predicted output vector in a prediction time domain according to the second discrete state update equation, and determining a driver reference trajectory vector and a driving system reference trajectory vector; constructing the driver trajectory cost function and the driving system trajectory cost function respectively by using the predicted output vector, the driver reference trajectory vector and the driving system reference trajectory vector to obtain the human-machine path tracking control game model.

3. The method according to claim 1, wherein, The step of solving the human-machine path tracking control game model to obtain human-machine torque conflict information comprises: determining a recursive relationship of steering control value functions corresponding to the driver and the driving system respectively under a Nash equilibrium condition by using a stochastic dynamic programming algorithm; calculating a closed-loop Nash equilibrium solution corresponding to the driver and the driving system respectively as the human-machine torque conflict information based on the first discrete state update equation and the recursive relationship.

4. The method of claim 1, wherein, The step of solving the human-machine path tracking control game model to obtain human-machine torque conflict information comprises: determining a recursive relationship of steering control value functions corresponding to the driver and the driving system respectively under a Stackelberg equilibrium condition by using a stochastic dynamic programming algorithm; determining a driver reaction function according to the recursive relationship of the steering control value functions corresponding to the driving system; calculating a corresponding open-loop Stackelberg equilibrium solution of the driver based on the corresponding open-loop Stackelberg equilibrium solution of the driving system; calculating a corresponding open-loop Stackelberg equilibrium solution of the driver based on the corresponding open-loop Stackelberg equilibrium solution of the driving system, as the human-machine torque conflict information.

5. The method according to claim 2, wherein, The solving the human-machine path tracking control game model to obtain human-machine torque conflict information comprises: solving a model closed-form solution corresponding to the human-machine path tracking control game model to obtain a relationship expression between human-machine steering control and target trajectory based on the model closed-form solution; solving the relationship expression by using a convex iteration algorithm to obtain corresponding open-loop Nash equilibrium solutions of the driver and the driving system as the human-machine torque conflict information.

6. The method according to claim 2, wherein, The solving the human-machine path tracking control game model to obtain human-machine torque conflict information comprises: converting the driving system trajectory cost function into a driving system trajectory optimization function considering the driver reaction function; solving the driving system trajectory optimization function to obtain a corresponding open-loop Stackelberg equilibrium solution of the driving system; calculating a corresponding open-loop Stackelberg equilibrium solution of the driver based on the corresponding open-loop Stackelberg equilibrium solution of the driving system, as the human-machine torque conflict information.

7. A human-machine co-pilot conflict control device characterized by comprising: comprises: a modeling module configured to establish a human-machine path tracking control game model corresponding to human-machine interaction behavior based on a deterministic steering torque of the driver and a random steering torque of the driver, wherein the human-machine path tracking control game model comprises a driver trajectory cost function and a driving system trajectory cost function; a solving module configured to solve the human-machine path tracking control game model to obtain human-machine torque conflict information; an application module configured to determine a shared control strategy based on the human-machine torque conflict information, and to control a vehicle based on the shared control strategy; when the human-machine path tracking control game model is a closed-loop game model, the modeling module is configured to establish the human-machine path tracking control game model corresponding to human-machine interaction behavior based on the deterministic steering torque of the driver and the random steering torque of the driver, comprising: establishing a first discrete state update equation of a human-machine co-driving vehicle dynamics system in a closed-loop information mode based on the deterministic steering torque of the driver and the random steering torque of the driver; augmenting the first discrete state update equation through a human-machine preview dynamic process to obtain a path tracking augmented system containing a human-machine preview state; constructing the driver trajectory cost function and the driving system trajectory cost function based on the path tracking augmented system to obtain the human-machine path tracking control game model; when the human-machine path tracking control game model is an open-loop game model, the driver trajectory cost function and the driving system trajectory cost function are as follows: wherein: , are weighting matrices of the path following control, is a model predicted output vector in a prediction horizon, , are reference trajectory vectors of the driver and the autonomous driving system, respectively, , are control input vectors of the driver and the autonomous driving system, respectively. 8.A computer readable storage medium having stored thereon a computer program, the program being executed by a processor to implement the human-machine co-driving conflict control method according to any one of claims 1 to 6.

9. An electronic device, comprising: comprises: one or more processors; A storage device is configured to store one or more programs, which when executed by the one or more processors, cause the one or more processors to implement the human-co-pilot conflict control method according to any one of claims 1 to 6.

Images