Vehicle emergency braking method and device, electronic equipment and readable storage medium

Abstract

The embodiment of the invention discloses a vehicle emergency braking method and device, electronic equipment and a readable storage medium, relates to the technical field of automatic driving, and aims to adjust the triggering opportunity of emergency braking in a personalized manner to a certain extent. The method comprises the following steps: acquiring current speed information of a current vehicle, current speed information of a first target vehicle and current distance information between the current vehicle and the first target vehicle; according to the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and a preset emergency braking triggering condition, whether emergency braking needs to be triggered or not is judged; wherein the preset emergency braking triggering condition is determined based on behavior characteristics of a driver of the current vehicle; if yes, an emergency braking starting instruction is sent to a controller of the current vehicle so as to brake the current vehicle. The brake device is suitable for braking the vehicle.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a vehicle emergency braking method, device, electronic device and readable storage medium. Background Art

[0002] The Autonomous Emergency Braking system (AEB) is an active safety technology for automobiles. It detects the distance to the target ahead through radar and smart cameras, analyzes it through an electronic control unit, determines whether there is a risk of collision based on different distances and speeds, and issues an alarm to the driver. If the driver does not brake in time after the alarm is displayed, the AEB system will be activated when the collision time is less than the time (trigger condition or trigger timing), automatically braking or slowing down the vehicle to the maximum extent, thereby reducing the probability of collision with the target ahead and avoiding accidents.

[0003] However, existing automatic emergency braking systems usually preset a fixed set of braking strategies and cannot adapt to the habits and preferences of different drivers. Summary of the invention

[0004] In view of this, the embodiments of the present application provide a vehicle emergency braking method, device, electronic device and readable storage medium, which can, to a certain extent, personalize the triggering timing of emergency braking.

[0005] In a first aspect, an embodiment of the present application provides a vehicle emergency braking method, comprising: obtaining current speed information of a current vehicle, current speed information of a first target vehicle, and current distance information between the current vehicle and the first target vehicle; judging whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information, and a preset emergency braking trigger condition; wherein the preset emergency braking trigger condition is determined based on the behavioral characteristics of the driver of the current vehicle; if necessary, sending an emergency braking start instruction to the controller of the current vehicle to brake the current vehicle.

[0006] According to a specific implementation method of an embodiment of the present application, before judging whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking trigger condition, the method also includes: determining the behavioral characteristics of the driver of the current vehicle; wherein, determining the behavioral characteristics of the driver of the current vehicle includes: based on a reinforcement learning function and preset samples, learning the behavioral characteristics of the driver of the current vehicle to obtain a performance parameter vector corresponding to the controller of the current vehicle; wherein different performance parameter vectors correspond to different behavioral characteristics; wherein each sample in the preset samples includes the speed information of the current vehicle, the speed information of the second target vehicle and the distance between the current vehicle and the second target vehicle; determining the behavioral characteristics of the driver of the current vehicle based on the performance parameter vector and the performance parameter vector corresponding to the preset multiple behavioral characteristics.

[0007] According to a specific implementation manner of an embodiment of the present application, before learning the behavioral characteristics of the driver of the current vehicle based on a reinforcement learning function and preset samples, the method also includes: determining a reinforcement learning function; wherein, determining the reinforcement learning function includes: establishing a following model of the current vehicle; wherein, the following model includes a continuous state-space equation; determining a control law of a controller of the current vehicle based on a performance index function of the current vehicle; wherein, the control law is an equation of an input quantity of the controller and a performance parameter vector of the controller; determining a reinforcement learning function based on the continuous state-space equation, the control law, and a state-action value function based on a reinforcement learning method.

[0008] According to a specific implementation method of an embodiment of the present application, determining the control law of the controller of the current vehicle based on the performance index function of the current vehicle includes: constructing a target discrete Hamiltonian equation and a target Riccati equation based on the performance index function of the current vehicle and the Hamiltonian equation; determining the control law of the controller of the current vehicle based on the target discrete Hamiltonian equation and the target Riccati equation.

[0009] According to a specific implementation of the embodiment of the present application, the control law of the controller of the current vehicle is determined according to the following formula:

[0010]

[0011] in, for The input quantity of the controller at the moment; is the first gain of the controller of the current vehicle, is a second gain of the controller of the current vehicle; is a positive semidefinite matrix, is a positive definite symmetric matrix, and represents a first performance parameter of the controller of the current vehicle, A second performance parameter representing a controller of the current vehicle; for Moment state quantity; is the first coefficient in the continuous state space equation, a second coefficient of the continuous state-space equation; is the first Hamiltonian matrix in the target discrete Hamiltonian equation, is the second Hamiltonian matrix in the target discrete Hamiltonian equation; is the disturbance in the continuous state space equation.

[0012] According to a specific implementation of the embodiment of the present application, the reinforcement learning function is determined according to the following formula:

[0013]

[0014] in, is the reinforcement learning function; is the weight vector; is the basis function.

[0015] According to a specific implementation method of an embodiment of the present application, the behavior characteristics of the driver of the current vehicle are learned based on a reinforcement learning function and a preset sample to obtain a performance parameter vector corresponding to the controller of the current vehicle, including: initializing the weight vector and the number of iterations in the reinforcement learning function to obtain an initial weight vector and setting the number of iterations to 1; constructing the Bellman equation in reinforcement learning according to the Bellman optimality principle and the reinforcement learning function; obtaining a new weight vector according to the preset sample, the number of iterations, the Bellman equation in reinforcement learning and the initial weight vector; if the difference between the new weight vector and the initial weight vector is less than or equal to a preset threshold, taking the new weight vector as the final weight vector; determining the performance parameter vector corresponding to the controller of the current vehicle according to the final weight vector and the reinforcement learning function; if the difference between the new weight vector and the initial weight vector is greater than the preset threshold, adding 1 to the number of iterations, and iterating cyclically until the difference between the new weight vector and the initial weight vector is less than or equal to the preset threshold.

[0016] According to a specific implementation method of an embodiment of the present application, the speed information includes speed and acceleration; the preset emergency braking trigger condition is a collision time threshold; wherein, judging whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking trigger condition, includes: determining the collision time from the current moment to the collision based on the current speed of the current vehicle, the current acceleration of the current vehicle, the current speed of the first target vehicle, the current acceleration of the first target vehicle and the current distance information; if the collision time is less than or equal to the collision time threshold, emergency braking needs to be triggered.

[0017] According to a specific implementation manner of an embodiment of the present application, before obtaining the current speed information of the current vehicle, the current speed information of the first target vehicle, and the current distance information between the current vehicle and the first target vehicle, the method also includes: determining a plurality of vehicles corresponding to the current vehicle and within a preset range; respectively obtaining the speed information of the plurality of vehicles, and the distance information between each of the plurality of vehicles and the current vehicle; determining the threat coefficient corresponding to each of the plurality of vehicles based on the speed information of the current vehicle, the speed information of each of the plurality of vehicles, and the distance information between each of the plurality of vehicles and the current vehicle; and determining the vehicle with the largest threat coefficient among the plurality of vehicles as the first target vehicle.

[0018] In a second aspect, an embodiment of the present application provides a vehicle emergency braking device, comprising: an acquisition module, used to acquire current speed information of a current vehicle, current speed information of a first target vehicle, and current distance information between the current vehicle and the first target vehicle; a judgment module, used to judge whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information, and a preset emergency braking trigger condition; wherein the preset emergency braking trigger condition is determined based on the behavioral characteristics of the driver of the current vehicle; a sending module, used to send an emergency braking start instruction to the controller of the current vehicle, if necessary, to brake the current vehicle.

[0019] According to a specific implementation method of an embodiment of the present application, the device also includes: a first determination module, which is used to determine the behavioral characteristics of the driver of the current vehicle before the judgment module determines whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking trigger condition; wherein, the first determination module includes: an acquisition submodule, which is used to learn the behavioral characteristics of the driver of the current vehicle based on a reinforcement learning function and preset samples, and obtain a performance parameter vector corresponding to the controller of the current vehicle; wherein different performance parameter vectors correspond to different behavioral characteristics; wherein each sample in the preset samples includes the speed information of the current vehicle, the speed information of the second target vehicle and the distance between the current vehicle and the second target vehicle; a determination submodule, which is used to determine the behavioral characteristics of the driver of the current vehicle based on the performance parameter vector and the performance parameter vector corresponding to the preset multiple behavioral characteristics.

[0020] According to a specific implementation method of an embodiment of the present application, the device also includes: a second determination module, which is used to determine the reinforcement learning function before the obtaining submodule learns the behavioral characteristics of the driver of the current vehicle based on the reinforcement learning function and preset samples; wherein, the second determination module includes: an establishment submodule, which is used to establish a following model of the current vehicle; wherein, the following model includes a continuous state space equation; a first determination submodule, which is used to determine the control law of the controller of the current vehicle according to the performance index function of the current vehicle; wherein, the control law is an equation of the input quantity of the controller and the performance parameter vector of the controller; a second determination submodule, which is used to determine the reinforcement learning function according to the continuous state space equation, the control law and the state-action value function based on the reinforcement learning method.

[0021] According to a specific implementation method of an embodiment of the present application, the first determination submodule is specifically used to: construct a target discrete Hamiltonian equation and a target Riccati equation based on the performance index function and the Hamiltonian equation of the current vehicle; and determine the control law of the controller of the current vehicle based on the target discrete Hamiltonian equation and the target Riccati equation.

[0022] According to a specific implementation of the embodiment of the present application, the first determination submodule is specifically used to determine the control law of the controller of the current vehicle according to the following formula:

[0023]

[0024] in, for The input quantity of the controller at the moment; is the first gain of the controller of the current vehicle, is a second gain of the controller of the current vehicle; is a positive semidefinite matrix, is a positive definite symmetric matrix, and represents a first performance parameter of the controller of the current vehicle, A second performance parameter representing a controller of the current vehicle; for Moment state quantity; is the first coefficient in the continuous state space equation, a second coefficient of the continuous state-space equation; is the first Hamiltonian matrix in the target discrete Hamiltonian equation, is the second Hamiltonian matrix in the target discrete Hamiltonian equation; is the disturbance in the continuous state space equation.

[0025] According to a specific implementation of the embodiment of the present application, the second determination submodule is specifically used to determine the reinforcement learning function according to the following formula:

[0026]

[0027] in, is the reinforcement learning function; is the weight vector; is the basis function.

[0028] According to a specific implementation method of an embodiment of the present application, the obtaining submodule is specifically used to initialize the weight vector and the number of iterations in the reinforcement learning function, obtain the initial weight vector and set the number of iterations to 1; according to the Bellman optimality principle and the reinforcement learning function, construct the Bellman equation in reinforcement learning; according to the preset sample, the number of iterations, the Bellman equation in reinforcement learning and the initial weight vector, obtain a new weight vector; if the difference between the new weight vector and the initial weight vector is less than or equal to a preset threshold, use the new weight vector as the final weight vector; according to the final weight vector and the reinforcement learning function, determine the performance parameter vector corresponding to the controller of the current vehicle; if the difference between the new weight vector and the initial weight vector is greater than the preset threshold, add 1 to the number of iterations, and iterate until the difference between the new weight vector and the initial weight vector is less than or equal to the preset threshold.

[0029] According to a specific implementation method of an embodiment of the present application, the speed information includes speed and acceleration; the preset emergency braking trigger condition is a collision time threshold; wherein the judgment module is specifically used to: determine the collision time at the current moment based on the current speed of the current vehicle, the current acceleration of the current vehicle, the current speed of the first target vehicle, the current acceleration of the first target vehicle and the current distance information; if the collision time is less than or equal to the collision time threshold, emergency braking needs to be triggered.

[0030] According to a specific implementation method of an embodiment of the present application, the device is also used to: determine a plurality of vehicles corresponding to the current vehicle and within a preset range before the acquisition module acquires the current speed information of the current vehicle, the current speed information of the first target vehicle, and the current distance information between the current vehicle and the first target vehicle; respectively acquire the speed information of the plurality of vehicles, and the distance information between each of the plurality of vehicles and the current vehicle; determine the threat coefficient corresponding to each of the plurality of vehicles based on the speed information of the current vehicle, the speed information of each of the plurality of vehicles, and the distance information between each of the plurality of vehicles and the current vehicle; and determine the vehicle with the largest threat coefficient among the plurality of vehicles as the first target vehicle.

[0031] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a housing, a processor, a memory, a circuit board and a power supply circuit, wherein the circuit board is placed inside the space enclosed by the housing, and the processor and the memory are arranged on the circuit board; a power supply circuit is used to supply power to various circuits or devices of the above-mentioned electronic device; the memory is used to store executable program code; the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the vehicle emergency braking method described in any of the aforementioned implementation methods.

[0032] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the vehicle emergency braking method described in any of the aforementioned implementation methods.

[0033] The vehicle emergency braking method, device, electronic device and readable storage medium of the present embodiment determine whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking trigger condition. If necessary, an emergency braking start instruction is sent to the controller of the current vehicle to brake the current vehicle. Since the preset emergency braking trigger condition is determined based on the behavioral characteristics of the driver of the current vehicle, the determined emergency braking trigger condition is different for different behavioral characteristics of the driver. Therefore, to a certain extent, the triggering timing of emergency braking can be personalized to adapt to the driving styles of different drivers, thereby improving the safety of the current vehicle while improving the adaptability of the automatic braking system to different drivers. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 A schematic diagram of a process flow of a vehicle emergency braking method provided in one embodiment of the present application;

[0036] Figure 2 A schematic diagram of a process flow of a vehicle emergency braking method provided in a specific embodiment of the present application;

[0037] Figure 3 A schematic diagram of the structure of a vehicle emergency brake device provided in one embodiment of the present application;

[0038] Figure 4 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0039] The embodiments of the present application are described in detail below in conjunction with the accompanying drawings. It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0040] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0041] In order to enable those skilled in the art to better understand the technical concept, implementation plan and beneficial effects of the embodiments of the present application, specific embodiments are described in detail below.

[0042] Figure 1 A schematic diagram of a vehicle emergency braking method according to an embodiment of the present application is shown in FIG. Figure 1 As shown, the vehicle emergency braking method of this embodiment may include:

[0043] S101, obtaining current speed information of a current vehicle, current speed information of a first target vehicle, and current distance information between the current vehicle and the first target vehicle.

[0044] The first target vehicle may be a vehicle followed by the current vehicle, and the first target vehicle is in front of the current vehicle.

[0045] The current distance information between the current vehicle and the first target vehicle may be the distance between the front ends of the two vehicles, or may be the distance between the front end of the current vehicle and the rear end of the first target vehicle.

[0046] In actual applications, the current vehicle can obtain the above information in real time to provide judgment conditions for determining whether to trigger emergency braking.

[0047] The current speed information of the current vehicle can be obtained through the vehicle bus (CAN bus) such as the speed of the current vehicle and the acceleration (deceleration) and other speed information. In one embodiment, the deceleration can be determined by obtaining the torque of the engine and the braking deceleration of the current vehicle.

[0048] The relevant information of the first target vehicle can be obtained through the camera, sensor and other equipment installed on the current vehicle, and the obtained information is processed to obtain the current speed information of the first target vehicle and the current distance information between the current vehicle and the first target vehicle.

[0049] S102: Determine whether emergency braking needs to be triggered based on current speed information of the current vehicle, current speed information of the first target vehicle, current distance information, and preset emergency braking triggering conditions.

[0050] The preset emergency braking triggering condition is determined based on the behavioral characteristics of the driver of the current vehicle.

[0051] The driver's behavior characteristics in this embodiment correspond to the driver's driving style and can reflect the driver's driving habits and preferences.

[0052] The emergency brake triggering condition can also become the emergency brake triggering opportunity.

[0053] S103: If necessary, send an emergency braking start instruction to a controller of the current vehicle to brake the current vehicle.

[0054] In step S102, if the judgment result is that emergency braking needs to be triggered, that is, the current vehicle needs to be braked, an emergency braking start instruction is sent to the controller (control system) of the current vehicle, so that the controller of the current vehicle controls the underlying brake actuator of the current vehicle to brake.

[0055] Braking the current vehicle in this embodiment includes decelerating the current vehicle to a certain speed, which is sufficient to avoid collision with the first target vehicle. Braking the current vehicle may also include reducing the speed of the current vehicle to 0, ie, stopping the current vehicle.

[0056] In this embodiment, whether emergency braking needs to be triggered is determined based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking triggering conditions. If necessary, an emergency braking start instruction is sent to the controller of the current vehicle to brake the current vehicle. Since the preset emergency braking triggering conditions are determined based on the behavioral characteristics of the driver of the current vehicle, the determined emergency braking triggering conditions are different for different behavioral characteristics of the driver. Therefore, the triggering timing of emergency braking can be personalized to a certain extent to adapt to the driving styles of different drivers, thereby improving the safety of the current vehicle and the adaptability of the automatic braking system to different drivers.

[0057] There are many ways to determine the behavior characteristics of the driver of the current vehicle. As a preferred embodiment, before determining whether to trigger emergency braking based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking triggering condition in step S101, the method of this embodiment may also include:

[0058] S104: Determine the behavior characteristics of the driver of the current vehicle.

[0059] In this embodiment, determining the behavior characteristics of the driver of the current vehicle in step S104 may include:

[0060] S104a, based on the reinforcement learning function and the preset samples, the behavior characteristics of the driver of the current vehicle are learned to obtain a performance parameter vector corresponding to the controller of the current vehicle.

[0061] In this embodiment, different performance parameter vectors correspond to different behavior characteristics.

[0062] Each sample in the preset samples includes speed information of the current vehicle, speed information of the second target vehicle, and a distance between the current vehicle and the second target vehicle.

[0063] S104b. Determine the behavior characteristics of the driver of the current vehicle according to the performance parameter vector and the performance parameter vector corresponding to the preset multiple behavior characteristics.

[0064] In a specific example, the behavior characteristics, i.e., driving styles, include aggressive, robust, and conservative, and each behavior characteristic corresponds to a set of parameters (vector).

[0065] As a preferred embodiment, before learning the behavior characteristics of the driver of the current vehicle based on the reinforcement learning function and the preset samples, the method further includes:

[0066] S105. Determine a reinforcement learning function.

[0067] Wherein, determining the reinforcement learning function in step S105 may include:

[0068] S105a: Establish a following model for the current vehicle.

[0069] In this embodiment, the following model includes continuous state space equations.

[0070] The continuous state space equation includes state quantity, control quantity and disturbance quantity, wherein the state quantity is related to the relative distance between the current vehicle and the preceding vehicle, the acceleration of the current vehicle and the acceleration of the preceding vehicle; the control quantity can be the input quantity of the controller; and the disturbance quantity can be the acceleration of the preceding vehicle.

[0071] S105b. Determine a control law of a controller of the current vehicle according to a performance indicator function of the current vehicle.

[0072] The performance index function of the current vehicle is established based on the control quantity of the controller, the state quantity in the continuous state space equation and the performance parameters corresponding to the controller.

[0073] In this embodiment, the control law is an equation of the input quantity of the controller and the performance parameter vector of the controller.

[0074] S105c, determining a reinforcement learning function according to the continuous state space equation, the control law, and the state action value function based on the reinforcement learning method.

[0075] The state-action-value function in the reinforcement learning method is established based on the state quantity in the continuous state space equation and the reward in reinforcement learning.

[0076] In some examples, determining the control law of the controller of the current vehicle according to the performance indicator function of the current vehicle in step S105b may include:

[0077] A1. According to the performance index function and Hamiltonian equation of the current vehicle, the target discrete Hamiltonian equation and the target Riccati equation are constructed.

[0078] The Hamiltonian equation is an equation in the prior art.

[0079] The target discrete Hamiltonian equation and the target Riccati equation in this embodiment are obtained by optimizing the Hamiltonian equation in the prior art by combining the state quantities in the continuous state space equation of the current vehicle in this embodiment and the relevant parameters in the performance index function.

[0080] A2. Determine the control law of the controller of the current vehicle based on the target discrete Hamiltonian equation and the target Riccati equation.

[0081] In some examples, the reinforcement learning function is determined as follows:

[0082] 1. Establish a vehicle following model based on the requirements of the adaptive vehicle following problem and the longitudinal kinematics and dynamics characteristics of the vehicle;

[0083] In order to describe the longitudinal dynamic characteristics of the vehicle in the following scenario, the system introduces two state variables: , They are:

[0084]

[0085]

[0086] in, is the relative distance between the vehicle and the vehicle ahead, is the reference relative distance, , It is the safe parking distance. is the expected headway, is the vehicle speed, is the speed of the vehicle ahead; therefore, the longitudinal dynamics differential equation of the vehicle in the following scenario can be expressed using the above two state variables:

[0087]

[0088]

[0089]

[0090] in, is the longitudinal acceleration of the vehicle, is the longitudinal acceleration of the preceding vehicle, is the expected headway; it is assumed here that the inner loop acceleration tracking control module of the vehicle can be expanded by first-order approximation:

[0091] (3)

[0092] in, is the inner loop dynamic time constant, is the longitudinal acceleration of the vehicle, is the expected longitudinal acceleration.

[0093] In a continuous system, the state quantity is chosen as , the control quantity is , the front vehicle acceleration is modeled as a system disturbance, that is, the disturbance ; Therefore, the continuous state space equation of the vehicle following system can be expressed as:

[0094] (4)

[0095] in, , , ; is the longitudinal acceleration of the vehicle, is the expected longitudinal acceleration, is the inner loop dynamic time constant, is the expected headway.

[0096] In controller design, the zero-order hold method (ZOH) is used to obtain the discrete system expression of the state space equation of the vehicle following system, and the state vector of the discrete system is defined as , control input , so in At any time you can get:

[0097] (5)

[0098] in, , is the sampling time interval, is the expected headway, is the inner loop dynamic time constant.

[0099] 2 The optimal cruise control problem is established as a linear quadratic optimization problem. The specific method is as follows:

[0100] 2.1 Establishment of the optimal quadratic control problem.

[0101] Performance indicator function It can be expressed as:

[0102]

[0103] in, is a positive semidefinite symmetric matrix, is a positive definite symmetric matrix, for The state quantity at the moment, for Control the amount at all times, and The parameter matrix represents the controller's requirements for driving style, comfort, and economy, and is defined here as:

[0104]

[0105] in, , , and They are respectively the first parameter, the second parameter, the third parameter and the fourth parameter which characterize the performance of the controller.

[0106] Specifically:

[0107] This parameter is associated with the first state variable of the system, and it represents the penalty weight for the first state variable. If the first state variable is related to the vehicle speed error, then It represents the penalty weight for speed error.

[0108] This parameter is associated with the second state variable of the system, and it represents the penalty weight for the second state variable. If the second state variable is related to vehicle acceleration or headway error, then It represents the penalty weight for acceleration or headway error.

[0109] This parameter is associated with the third state variable of the system, which represents the penalty weight for the third state variable. If the third state variable is related to the acceleration rate of the vehicle or the headway rate of change, then It represents the penalty weight for the rate of change of acceleration or the rate of change of headway.

[0110] , ,and These values ​​reflect the importance that the system designer places on different control objectives. For example, if A larger value of means that the system focuses more on reducing the error associated with the second state variable, perhaps to improve ride comfort or ensure a safe distance.

[0111] is the rate of change of acceleration.

[0112] Substituting it into formula (6-1), the performance index function can also be expressed as:

[0113]

[0114] It can be clearly seen here that the penalty Characterizes the following performance, that is, when reaching the steady state, the vehicle maintains the desired following distance with the vehicle in front at the desired longitudinal speed. The penalty term Expected to minimize fuel consumption, penalty term It is expected to improve the comfort of the vehicle and reduce frequent acceleration and deceleration operations.

[0115] 2.2 Solution of the optimal quadratic control problem.

[0116] Define the discrete Hamiltonian equation to solve the above optimization problem. For solving the optimization problem (6-2), the discrete Hamiltonian equation is defined as (8)

[0117] in, is the Lagrange multiplier, The definition of is as described above and does not change in the derivation of the following formula According to the minimum principle, the Hamiltonian function The smallest optimal solution needs to satisfy:

[0118]

[0119]

[0120] because The inverse matrix Exists, from (9-1) we get:

[0121]

[0122] Here for Choose one of the following:

[0123] (11)

[0124] in, and is the Hamiltonian matrix. Substituting equations (10) and (11) into the state equation (4), we can obtain

[0125]

[0126] At the same time, substituting formula (11) into formula (9-2) we get:

[0127]

[0128] Substituting formula (12) into formula (13), we get:

[0129]

[0130]

[0131]

[0132] Since formula (14) requires all holds, and thus we get the Riccati equation:

[0133]

[0134] because The Riccati equation (15) is transformed into:

[0135]

[0136]

[0137] because In the current The time is unknown, without loss of generality, assuming , formula (16) becomes:

[0138] (17)

[0139] Therefore, about The explicit solution of is:

[0140] (18)

[0142] in, ; Therefore, combining equations (9-2), (10), and (11), the optimal control law is:

[0143]

[0144] for The input quantity of the controller at the moment; is the first gain of the controller of the current vehicle, is a second gain of the controller of the current vehicle; is a positive semidefinite matrix, is a positive definite symmetric matrix, and represents a first performance parameter of the controller of the current vehicle, A second performance parameter representing a controller of the current vehicle; for Moment state quantity; is the first coefficient in the continuous state space equation, a second coefficient of the continuous state-space equation; is the first Hamiltonian matrix in the target discrete Hamiltonian equation, is the second Hamiltonian matrix in the target discrete Hamiltonian equation; is the disturbance in the continuous state space equation.

[0145] in, is the controller gain associated with the current error in the system state variable, is the controller gain related to the rate of change of the system state variable. From this analysis, it can be seen that the driving style can be distinguished by changing the controller gain.

[0146] By adjusting and With these gains, the system can better balance control objectives such as safety, comfort, and economy. For example, for drivers who prefer an aggressive driving style, the system may reduce to reduce sensitivity to small speed changes; for drivers who prefer a conservative driving style, the system may increase value to respond faster to changes in the speed of the vehicle ahead.

[0147] 2.3 Acceleration estimation of the preceding vehicle considering uncertainty.

[0148] The front vehicle's speed is used to estimate the front vehicle's acceleration through two first-order low-pass filters, and finally a curvature limiter is used to limit abnormal acceleration changes.

[0149] 3. Self-learning module for following driving style

[0150] In view of the fact that the driving style of a specific driver may be different, the driving style of the specific driver is learned based on the reinforcement learning method, and then the control parameters of the optimal cruise control problem are adjusted to achieve the function of the self-learning system; the specific method is as follows:

[0151] 3.1 Definition of specific driver driving style.

[0152] Based on the analysis results of a large amount of driver statistical data, the usual driving style classification method is used to divide drivers into three categories: aggressive, steady, and conservative. Aggressive drivers tend to keep a close distance to the vehicle in front and accelerate and decelerate frequently, while conservative drivers tend to keep a long distance from the vehicle in front and accelerate and decelerate less frequently. The system uses the variable headway time commonly used in automatic emergency braking systems to characterize this behavioral style.

[0153]

[0154] in, is the headway time; is the rate of change of acceleration, , ,and The matrix of formula (7) is Elements of is the controller gain.

[0155] Smaller means aggressive drivers tend to keep a shorter distance to the car ahead; is large, indicating that this type of driver is likely to change acceleration frequently and drastically; , ,and The cost function or performance metric used to define the optimization problem quantifies the relative importance of different state variables in the control system design.

[0156] Represents the differential gain, which is a parameter in the controller used to adjust the system's response to the rate of change of the error. In a PID (proportional-integral-derivative) controller, It is the proportional constant of the differential term, which reduces the overshoot and oscillation of the system and improves the stability and response speed of the system. The value of can be adjusted according to the driver's driving style. For example, for a driver who prefers an aggressive driving style, may be set lower to reduce sensitivity to small speed changes, or higher for drivers who prefer a more conservative driving style. Might be set higher to respond more quickly to speed changes of the vehicle ahead.

[0157] 3.2 Establishment of reinforcement learning method.

[0158] Since the actual driver style information is unknown when the system is designed, the system uses reinforcement learning to learn the specific driver's driving style; when the Riccati equation is solved, reinforcement learning is completed and the system learns the driver's driving style habits; in the linear quadratic problem in the linear discrete system, The function is the state-action-value function in reinforcement learning, which is a quadratic form of state quantity and control quantity and can be expressed as:

[0159]

[0160] in, is the reward in reinforcement learning, is the Ricardo solution; Substituting the system model (5) into (20) we get

[0161]

[0162] in, , z(k) ＝ [ x T (k) , d T (k) , u T (k)] T For the convenience of subsequent differentiation, define the kernel matrix:

[0163]

[0164] According to Bellman's optimality principle, the optimal solution is The control amount of the minimum function value ,Right now:

[0165] Therefore, the derivative of equation (21) is get:

[0166]

[0167] From formula (21), we can get the system parameters For the kernel matrix The display expression of the kernel matrix is Solve using the least squares method in function approximation;

[0168] Use the linear kernel function Function approximation, which can be written as:

[0169]

[0170] in, is the reinforcement learning function; is the weight vector; is the basis function. Specifically, W ＝ [w1 ,w2 ,...,w15] T , .

[0171] Through incremental learning, the optimal value is gradually updated based on the driver data, and the 15 unknown parameters of the optimal weight vector are approximated and solved.

[0172] In some examples, the step S104a of learning the behavior characteristics of the driver of the current vehicle based on the reinforcement learning function and the preset samples to obtain a performance parameter vector corresponding to the controller of the current vehicle includes:

[0173] A1. Initialize the weight vector and the number of iterations in the reinforcement learning function to obtain an initial weight vector and set the number of iterations to 1.

[0174] The weight vector W (controller gains) is initialized. These initial values ​​can be random or set based on some prior knowledge.

[0175] A2. According to the Bellman optimality principle and the reinforcement learning function, the Bellman equation in reinforcement learning is constructed.

[0176] Using the parametric approximation form Function, for infinite time domain ,in, It is a value between 0 and 1 that is used to discount future returns to reflect the present value of future rewards. When the discount factor (γ) is equal to 1, it means that all future returns are given the same value as current returns, that is, each unit of future returns is equivalent to one unit of current returns.

[0177] based on , the Bellman equation in reinforcement learning can be written as:

[0178]

[0179] The optimal weight vector W is found by solving the Bellman equation.

[0180] A3. Obtain a new weight vector according to a preset sample, the number of iterations, the Bellman equation in the reinforcement learning and the initial weight vector.

[0181] The solution of equation (26) is transformed into the least squares method: Problem solving: Optimize the weight vector W through the least squares method so that the predicted value of the Q function is as close to the actual return value as possible.

[0182] Determine the number of batch training samples N (N is an integer greater than or equal to 1), ( The sample at the time (greater than or equal to 0, less than or equal to the time interval of sample collection) is Iteration calculation ( is the number of iterations, which is less than or equal to the preset total number of iterations), Is an integer constant, calculate in sequence:

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191] get The least squares estimate of is:

[0192]

[0193] in, and Respectively represent the system status in two consecutive short periods of time; Indicates that between two consecutive time steps, the function The amount of state change after processing is , the system can learn and optimize control parameters such as and , in order to achieve better control effect.

[0194] represents the reward value under the system state z(k) at time step k. In adaptive cruise control systems, it is used to evaluate the pros and cons of the current state and guide the system to take actions that maximize the cumulative reward. By optimizing the reward function, the system can learn and adapt to different driving styles and traffic conditions, improving driving safety and comfort.

[0195] In addition, the system collects data during actual driving, including the state of the vehicle (such as speed, acceleration, distance to the vehicle in front, etc.) and the actions taken by the driver (such as acceleration, deceleration, maintaining speed, etc.) as samples. In some examples, each sample includes the state of the current vehicle, the state of the vehicle in front, and the action taken by the driver of the current vehicle.

[0196] A4. If the difference between the new weight vector and the initial weight vector is less than or equal to a preset threshold, the new weight vector is used as the final weight vector.

[0197] A5. Determine a performance parameter vector corresponding to the controller of the current vehicle based on the final weight vector and the reinforcement learning function.

[0198] After the LS algorithm converges, the system compares the calculated performance parameter vector with the two norms of the three types of driver style gain parameter vectors given in the driver parameter table of different following behavior styles, and selects the driving style type with the smallest two norm value as the driving style of the current driver. According to the collision time corresponding to the current driver's driving style, the controller parameters are updated; otherwise, the original controller parameters are maintained and the next step of learning is continued.

[0199] A6. If the difference between the new weight vector and the initial weight vector is greater than the preset threshold, the number of iterations is increased by 1, and the iteration is repeated until the difference between the new weight vector and the initial weight vector is less than or equal to the preset threshold.

[0200] This embodiment uses a method that combines a linear quadratic control method and a reinforcement learning online learning method to better optimize system performance.

[0201] In some examples, the speed information includes speed and acceleration; and the preset emergency braking trigger condition is a collision occurrence time threshold.

[0202] Wherein, judging whether emergency braking needs to be triggered according to the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking triggering condition in step S102 may include: S102a, determining a collision time at a current moment from a collision to a collision according to a current speed of the current vehicle, a current acceleration of the current vehicle, a current speed of the first target vehicle, a current acceleration of the first target vehicle, and current distance information.

[0203] The collision occurrence time of the collision represents the time interval from the current moment to the collision between the current vehicle and the first target vehicle according to the current driving state.

[0204] The collision occurrence time in this embodiment can be determined by referring to the calculation method of calculating the collision occurrence time in the prior art.

[0205] S102b: If the collision occurrence time is less than or equal to the collision occurrence time threshold, emergency braking needs to be triggered.

[0206] When the collision occurrence time is less than or equal to the collision occurrence time threshold, it indicates that there is a risk of collision between the current vehicle and the first target vehicle, and emergency braking of the current vehicle is required.

[0207] In actual scenarios, multiple vehicles may appear in front of the current vehicle, but not every vehicle needs to be the target vehicle of the current vehicle. To reduce the risk of collision, in some examples, before obtaining the current speed information of the current vehicle, the current speed information of the first target vehicle, and the current distance information between the current vehicle and the first target vehicle in step S101, the method of this embodiment may further include: S106: Determine a plurality of vehicles corresponding to the current vehicle and within a preset range.

[0208] The plurality of vehicles may be two, three, five, etc. vehicles.

[0209] S107, respectively acquiring speed information of a plurality of vehicles and distance information between each of the plurality of vehicles and the current vehicle.

[0210] In some examples, the speed information may be speed and acceleration, and the speed may be the longitudinal speed of the vehicle.

[0211] The distance information may be the distance between the front of each vehicle and the front of the current vehicle, or the distance between the rear of each vehicle and the front of the current vehicle.

[0212] S108. Determine a threat coefficient corresponding to each of the multiple vehicles according to speed information of the current vehicle, speed information of each of the multiple vehicles, and distance information between each of the multiple vehicles and the current vehicle.

[0213] In some examples, the collision time when each of the multiple vehicles collides with the current vehicle may be calculated, and the collision time may be used as the threat coefficient corresponding to each vehicle.

[0214] S109: Determine the vehicle with the greatest threat coefficient among the multiple vehicles as the first target vehicle.

[0215] In some examples, the vehicle with the longest collision occurrence time among multiple vehicles may be determined as the first target vehicle, and then step S101 is performed.

[0216] The solution of the present application is described in detail below with a specific embodiment.

[0217] In one embodiment, an automatic emergency braking system with driving style learning capability includes a perception fusion module A, a driving style (driving behavior) self-learning module B, an emergency braking module C and a vehicle braking execution module D.

[0218] Among them, the perception fusion module A is used to obtain the driving status information of the vehicle in front, including the speed information of the vehicle in front and the relative distance to the current vehicle.

[0219] The driving style self-learning module B is used to learn the driver's driving style based on the reinforcement learning method, and then adjust the control parameters of the optimal cruise control problem to achieve the self-learning system function.

[0220] The emergency braking module C includes a target threat calculation submodule Ca, a deceleration calculation submodule Cb and a braking control submodule Cc; wherein the target threat calculation submodule Ca is used to filter surrounding targets and calculate target threats, and at the same time confirm the target with the largest current threat coefficient; the deceleration calculation submodule Cb is used to calculate the deceleration required for the target to brake to a stop; the braking control submodule Cc is used to determine the braking control parameters based on the calculated deceleration.

[0221] The vehicle braking execution module D is used to perform emergency braking on the current vehicle according to the braking control parameters.

[0222] See also Figure 2 The vehicle emergency braking method of this embodiment may include: 1. The perception fusion module A determines the speed information of the current vehicle and the preceding vehicle (the first target vehicle) as well as the distance between the current vehicle and the preceding vehicle.

[0223] 2. Determine whether the braking system needs braking.

[0224] 3. If the braking system does not need braking, the driving style self-learning module B evaluates whether the driver's driving style is consistent; if not, the driving style of the driver is learned to obtain the performance parameters of the controller; and whether convergence is determined; if not, the module returns to step 1; if convergence is achieved, the 2-norm of the calculated performance parameters and the performance parameters of the controller corresponding to different preset driving styles is calculated, and the driving style corresponding to the smallest 2-norm within the preset range is determined as the final driving style.

[0225] 4. If the braking system needs to brake, the emergency braking module C calculates the target that poses the greatest threat, obtains the speed information of the target, the speed information of the current vehicle, and the distance between the current vehicle and the target vehicle, and determines whether the trigger braking conditions are met; if met, the deceleration is calculated, and the braking control parameters are determined based on the deceleration.

[0226] 5. The vehicle braking execution module D brakes the current vehicle according to the braking control parameters.

[0227] The automatic braking system of this embodiment is suitable for L2 level autonomous driving vehicles. Its goal is to achieve automatic braking in emergency situations in actual driving scenarios through environmental perception and vehicle body information fusion. At the same time, the system takes into account the driving style of specific drivers, so that the system has consistent satisfaction and acceptance for different groups of people.

[0228] In some specific examples, the driver's driving style (driving behavior) is learned based on a reinforcement learning algorithm.

[0229] The specific algorithm process is: Step (1): Initialize the weight vector based on the controller gain in the driver parameter table for different following behavior styles And the characteristic parameters of each driving style, such as the parameters in Table 1.

[0230] Step (2): Update and calculate the current data set: ; Step (3): Use LS algorithm: weight vector Update using formula (27) to get ; Step (4): If the LS algorithm converges, we get , and then update the kernel matrix: using the weight vector Inverse kernel matrix Each parameter, get the controller gain estimate The evaluation parameters are compared with the two norms of the three driving style parameter vectors and the most suitable set of parameters in the driver parameter table with different following behavior styles is selected. Update controller parameters, set , and end the calculation; and return the vehicle to follow the driving style; Step (5): If the LS algorithm does not converge, set ; Continue with steps (2)-(4).

[0231] Figure 3 A schematic diagram of the structure of a vehicle emergency brake device provided in an embodiment of the present application is shown in FIG. Figure 3As shown, the vehicle emergency braking device of this embodiment includes: an acquisition module 11, used to acquire the current speed information of the current vehicle, the current speed information of the first target vehicle and the current distance information between the current vehicle and the first target vehicle; a judgment module 12, used to judge whether it is necessary to trigger emergency braking according to the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking triggering condition; wherein the preset emergency braking triggering condition is determined based on the behavioral characteristics of the driver of the current vehicle; a sending module 13, used to send an emergency braking start instruction to the controller of the current vehicle to brake the current vehicle if necessary.

[0232] The device of this embodiment can be used to perform Figure 1 The technical solution of the method embodiment shown has similar implementation principles and technical effects, which will not be repeated here.

[0233] The device of this embodiment determines whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking triggering conditions. If necessary, a start emergency braking instruction is sent to the controller of the current vehicle to brake the current vehicle. Since the preset emergency braking triggering conditions are determined based on the behavioral characteristics of the driver of the current vehicle, the determined emergency braking triggering conditions are different for different behavioral characteristics of the driver. Therefore, to a certain extent, the triggering timing of emergency braking can be personalized to adapt to the driving styles of different drivers, thereby improving the safety of the current vehicle while improving the adaptability of the automatic braking system to different drivers.

[0234] As an optional embodiment, the device also includes: a first determination module, which is used to determine the behavioral characteristics of the driver of the current vehicle before the judgment module determines whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking trigger condition; wherein, the first determination module includes: an acquisition submodule, which is used to learn the behavioral characteristics of the driver of the current vehicle based on a reinforcement learning function and preset samples, and obtain a performance parameter vector corresponding to the controller of the current vehicle; wherein different performance parameter vectors correspond to different behavioral characteristics; wherein each sample in the preset samples includes the speed information of the current vehicle, the speed information of the second target vehicle and the distance between the current vehicle and the second target vehicle; a determination submodule, which is used to determine the behavioral characteristics of the driver of the current vehicle based on the performance parameter vector and the performance parameter vector corresponding to the preset multiple behavioral characteristics.

[0235] As an optional embodiment, the device also includes: a second determination module, used to determine the reinforcement learning function before the obtaining submodule learns the behavioral characteristics of the driver of the current vehicle based on the reinforcement learning function and preset samples; wherein, the second determination module includes: an establishment submodule, used to establish a following model of the current vehicle; wherein, the following model includes a continuous state space equation; a first determination submodule, used to determine the control law of the controller of the current vehicle according to the performance index function of the current vehicle; wherein, the control law is an equation of the input quantity of the controller and the performance parameter vector of the controller; the second determination submodule, used to determine the reinforcement learning function according to the continuous state space equation, the control law and the state action value function based on the reinforcement learning method.

[0236] As an optional implementation, the first determination submodule is specifically used to: construct a target discrete Hamiltonian equation and a target Riccati equation based on the performance index function and the Hamiltonian equation of the current vehicle; and determine the control law of the controller of the current vehicle based on the target discrete Hamiltonian equation and the target Riccati equation.

[0237] As an optional implementation manner, the first determination submodule is specifically configured to determine the control law of the controller of the current vehicle according to the following formula:

[0238] Among them, u(k) is the input of the controller at time k; kx is the first gain of the controller of the current vehicle, kd is the second gain of the controller of the current vehicle; Q is a semi-positive definite matrix, R is a positive definite symmetric matrix, and Q represents the first performance parameter of the controller of the current vehicle, and R represents the second performance parameter of the controller of the current vehicle; x(k) is the state quantity at time k; A is the first coefficient in the continuous state space equation, and Bu is the second coefficient in the continuous state space equation; P is the first Hamiltonian matrix in the target discrete Hamiltonian equation, and h is the second Hamiltonian matrix in the target discrete Hamiltonian equation; d(k) is the disturbance quantity in the continuous state space equation.

[0239] As an optional implementation manner, the second determination submodule is specifically configured to determine the reinforcement learning function according to the following formula:

[0240] in, is the reinforcement learning function; W is the weight vector; is the basis function.

[0241] The weight vector and the number of iterations in the learning function are used to obtain an initial weight vector and the number of iterations is set to 1; according to the Bellman optimality principle and the reinforcement learning function, the Bellman equation in reinforcement learning is constructed; according to the preset sample, the number of iterations, the Bellman equation in reinforcement learning and the initial weight vector, a new weight vector is obtained; if the difference between the new weight vector and the initial weight vector is less than or equal to a preset threshold, the new weight vector is used as the final weight vector; according to the final weight vector and the reinforcement learning function, a performance parameter vector corresponding to the controller of the current vehicle is determined; if the difference between the new weight vector and the initial weight vector is greater than the preset threshold, the number of iterations is increased by 1, and the iteration is repeated until the difference between the new weight vector and the initial weight vector is less than or equal to the preset threshold.

[0242] As an optional implementation, the speed information includes speed and acceleration; the preset emergency braking trigger condition is a collision time threshold; wherein the judgment module is specifically used to: determine the collision time at the current moment based on the current speed of the current vehicle, the current acceleration of the current vehicle, the current speed of the first target vehicle, the current acceleration of the first target vehicle and the current distance information; if the collision time is less than or equal to the collision time threshold, emergency braking needs to be triggered.

[0243] As an optional embodiment, the device is also used to: determine a plurality of vehicles corresponding to the current vehicle and within a preset range before the acquisition module acquires the current speed information of the current vehicle, the current speed information of the first target vehicle, and the current distance information between the current vehicle and the first target vehicle; respectively acquire the speed information of the plurality of vehicles, and the distance information between each of the plurality of vehicles and the current vehicle; determine the threat coefficient corresponding to each of the plurality of vehicles based on the speed information of the current vehicle, the speed information of each of the plurality of vehicles, and the distance information between each of the plurality of vehicles and the current vehicle; and determine the vehicle with the largest threat coefficient among the plurality of vehicles as the first target vehicle.

[0244] The device of the above embodiment can be used to execute the technical solution of the above method embodiment. Its implementation principle and technical effect are similar and will not be repeated here.

[0245] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown in FIG. Figure 4As shown, it may include: a shell 41, a processor 42, a memory 43, a circuit board 44 and a power supply circuit 45, wherein the circuit board 44 is arranged inside the space enclosed by the shell 41, and the processor 42 and the memory 43 are arranged on the circuit board 44; the power supply circuit 45 is used to supply power to various circuits or devices of the above-mentioned electronic equipment; the memory 43 is used to store executable program codes; the processor 42 runs the program corresponding to the executable program code by reading the executable program code stored in the memory 43, so as to execute any one of the vehicle emergency braking methods provided in the aforementioned embodiments, and thus can also achieve corresponding beneficial technical effects, which have been described in detail in the previous text and will not be repeated here.

[0246] The above electronic devices exist in many forms, including but not limited to: (1) Mobile communication devices: These devices are characterized by their mobile communication functions and their main purpose is to provide voice and data communications. These terminals include: smart phones (such as iPhone), multimedia phones, feature phones, and low-end phones.

[0247] (2) Ultra-mobile personal computer devices: These devices fall into the category of personal computers and have computing and processing capabilities, and generally also have mobile Internet access features. These terminals include: PDAs, MIDs, and UMPC devices, such as the iPad.

[0248] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players (such as iPods), handheld game consoles, e-books, as well as smart toys and portable car navigation devices.

[0249] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server is similar to a general-purpose computer architecture, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0250] (5) Other electronic devices with data interaction functions.

[0251] Correspondingly, an embodiment of the present application also provides a computer-readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement any translation method provided by the aforementioned embodiments, thereby also being able to achieve the corresponding technical effects, which have been described in detail above and will not be repeated here.

[0252] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0253] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0254] In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0255] For the convenience of description, the above device is described by dividing the functions into various units / modules. Of course, when implementing the present application, the functions of each unit / module can be implemented in the same or multiple software and / or hardware.

[0256] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0257] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A vehicle emergency braking method, characterized in that: include: Acquire current speed information of the current vehicle, current speed information of the first target vehicle, and current distance information between the current vehicle and the first target vehicle; Determining whether emergency braking needs to be triggered according to the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and a preset emergency braking triggering condition; wherein the preset emergency braking triggering condition is determined based on the behavioral characteristics of the driver of the current vehicle; If necessary, an emergency braking instruction is sent to a controller of the current vehicle to brake the current vehicle.

2. The method according to claim 1, characterized in that Before determining whether emergency braking needs to be triggered based on the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information, and the preset emergency braking triggering condition, the method further includes: determining a behavioral characteristic of a driver of the current vehicle; Wherein, determining the behavior characteristics of the driver of the current vehicle includes: Based on the reinforcement learning function and the preset samples, the behavior characteristics of the driver of the current vehicle are learned to obtain a performance parameter vector corresponding to the controller of the current vehicle; wherein different performance parameter vectors correspond to different behavior characteristics; wherein each sample in the preset samples includes speed information of the current vehicle, speed information of a second target vehicle, and a distance between the current vehicle and the second target vehicle; The behavior characteristics of the driver of the current vehicle are determined according to the performance parameter vector and the performance parameter vectors corresponding to the preset multiple behavior characteristics.

3. The method according to claim 2, characterized in that Before learning the behavior characteristics of the driver of the current vehicle based on the reinforcement learning function and the preset samples, the method further includes: Determine the reinforcement learning function; Wherein, determining the reinforcement learning function includes: Establishing a following model of the current vehicle; wherein the following model includes a continuous state space equation; Determining a control law of a controller of the current vehicle according to the performance index function of the current vehicle; wherein the control law is an equation of an input quantity of the controller and a performance parameter vector of the controller; A reinforcement learning function is determined according to the continuous state space equation, the control law and a state action value function based on a reinforcement learning method.

4. The method according to claim 3, characterized in that Determining the control law of the controller of the current vehicle according to the performance indicator function of the current vehicle includes: According to the performance index function and Hamiltonian equation of the current vehicle, construct a target discrete Hamiltonian equation and a target Riccati equation; A control law of a controller of the current vehicle is determined according to the target discrete Hamiltonian equation and the target Riccati equation.

5. The method according to claim 4, characterized in that The control law of the controller of the current vehicle is determined according to the following formula: ； in, for The input quantity of the controller at the moment; is the first gain of the controller of the current vehicle, is a second gain of the controller of the current vehicle; is a positive semidefinite matrix, is a positive definite symmetric matrix, and represents a first performance parameter of the controller of the current vehicle, A second performance parameter representing a controller of the current vehicle; for Moment state quantity; is the first coefficient in the continuous state space equation, The second coefficient of the continuous state space equation; is the first Hamiltonian matrix in the target discrete Hamiltonian equation, is the second Hamiltonian matrix in the target discrete Hamiltonian equation; is the disturbance in the continuous state space equation.

6. The method according to claim 5, characterized in that The reinforcement learning function is determined according to the following formula: ； in, is the reinforcement learning function; is the weight vector; is the basis function.

7. The method according to claim 6, characterized in that The learning of the behavior characteristics of the driver of the current vehicle based on the reinforcement learning function and the preset samples to obtain a performance parameter vector corresponding to the controller of the current vehicle includes: Initializing the weight vector and the number of iterations in the reinforcement learning function to obtain an initial weight vector and set the number of iterations to 1; According to the Bellman optimality principle and the reinforcement learning function, the Bellman equation in reinforcement learning is constructed; Obtaining a new weight vector according to a preset sample, the number of iterations, the Bellman equation in the reinforcement learning, and the initial weight vector; If the difference between the new weight vector and the initial weight vector is less than or equal to a preset threshold, the new weight vector is used as the final weight vector; Determining a performance parameter vector corresponding to the controller of the current vehicle according to the final weight vector and the reinforcement learning function; If the difference between the new weight vector and the initial weight vector is greater than the preset threshold, the number of iterations is increased by 1, and the iteration is repeated until the difference between the new weight vector and the initial weight vector is less than or equal to the preset threshold.

8. The method according to claim 1, characterized in that: The speed information includes speed and acceleration; the preset emergency braking trigger condition is a collision occurrence time threshold; Wherein, judging whether emergency braking needs to be triggered according to the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and the preset emergency braking triggering condition includes: Determining a collision occurrence time at a current moment from a collision occurrence time according to a current speed of the current vehicle, a current acceleration of the current vehicle, a current speed of the first target vehicle, a current acceleration of the first target vehicle and the current distance information; If the collision occurrence time is less than or equal to the collision occurrence time threshold, emergency braking needs to be triggered.

9. The method according to claim 1, characterized in that: Before acquiring the current speed information of the current vehicle, the current speed information of the first target vehicle, and the current distance information between the current vehicle and the first target vehicle, the method further includes: Determining a plurality of vehicles corresponding to the current vehicle and within a preset range; Respectively acquiring speed information of the plurality of vehicles and distance information between each of the plurality of vehicles and the current vehicle; Determine a threat coefficient corresponding to each of the multiple vehicles according to speed information of the current vehicle, speed information of each of the multiple vehicles, and distance information between each of the multiple vehicles and the current vehicle; The vehicle with the greatest threat coefficient among the multiple vehicles is determined as the first target vehicle.

10. A vehicle emergency braking device, characterized in that: include: An acquisition module, used to acquire current speed information of a current vehicle, current speed information of a first target vehicle, and current distance information between the current vehicle and the first target vehicle; A judgment module, used to judge whether emergency braking needs to be triggered according to the current speed information of the current vehicle, the current speed information of the first target vehicle, the current distance information and a preset emergency braking triggering condition; wherein the preset emergency braking triggering condition is determined based on the behavioral characteristics of the driver of the current vehicle; The sending module is used to send an emergency braking start instruction to the controller of the current vehicle to brake the current vehicle if necessary.

11. An electronic device, characterized in that: The electronic device comprises: a housing, a processor, a memory, a circuit board and a power supply circuit, wherein the circuit board is placed inside the space enclosed by the housing, and the processor and the memory are arranged on the circuit board; the power supply circuit is used to supply power to various circuits or devices of the above-mentioned electronic device; the memory is used to store executable program codes; the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the vehicle emergency braking method described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the vehicle emergency braking method described in any one of claims 1 to 9.

Images