Intelligent vehicle steering and braking coordinated collision avoidance control system and method

Abstract

The present invention discloses a smart car steering and braking coordinated collision avoidance control system and method, including a perception layer, a decision layer and an execution layer. The decision layer includes a lane change trajectory planning module, an expected maximum lateral acceleration output module, a coordinated collision avoidance expected braking deceleration output module, a coordinated collision avoidance critical safety distance calculation module and a collision avoidance strategy decision module. Based on the real-time status information of the vehicle and surrounding vehicles and the current road information, the vehicle's uniform lane change trajectory is preliminarily planned, and then the expected maximum lateral acceleration and the coordinated collision avoidance expected braking deceleration are decided and output to complete the coordinated collision avoidance trajectory planning; the coordinated collision avoidance critical safety distance is calculated based on the planned collision avoidance trajectory and driving information; the collision avoidance decision is completed by comprehensively considering the relationship between the current longitudinal distance between the vehicle and the obstacle and the critical safety distance and the traffic conditions of the adjacent lanes, and finally the corresponding operation is completed. The present invention comprehensively considers the influence of multiple safety factors to ensure the safety of vehicle driving and ride comfort.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle active safety, and in particular relates to a control system and method for coordinated collision avoidance of steering and braking of an intelligent vehicle. Background Art

[0002] As the number of motor vehicles in my country continues to rise, the demand for active vehicle safety technologies is also increasing. Active collision avoidance control systems, as a major category of active vehicle safety technologies, have become a key focus for major automakers and academics.

[0003] Currently, most active vehicle collision avoidance control systems focus on single braking and steering control. While research and application in these two areas are relatively mature, each still has its limitations. For example, when the vehicle's speed is too high, the braking distance required by the braking system is long. In this case, if an obstacle suddenly appears in the road ahead or the vehicle ahead suddenly brakes, the vehicle often cannot effectively avoid the collision due to the small distance between vehicles. While steering systems can address this issue, when the vehicle changes lanes by steering, the high speed often causes accidents such as rollover or rear-end collisions with vehicles in other lanes. Therefore, to achieve more effective collision avoidance control while fully considering the real reactions of human drivers in emergency situations, further research on coordinated steering and braking collision avoidance control systems is necessary. Clearly, relatively little research has been conducted in this area of ​​the prior art.

[0004] In the limited existing research on cooperative collision avoidance control, the maximum lateral acceleration is often determined solely by considering the ego vehicle's speed and the corresponding longitudinal displacement at the time of collision avoidance. However, in actual collision avoidance, drivers often make decisions based on their experience and subjective judgment of the current lane width, lane centerline, and the longitudinal distance between the ego vehicle and the obstacle. Clearly, existing technologies have given little consideration to this aspect, and therefore further refine the maximum lateral acceleration decision-making process to better reflect real-world driver behavior and enhance ride comfort. Furthermore, when determining the braking deceleration for cooperative collision avoidance, existing technologies often simply assume a specific value and use it to plan the vehicle's cooperative collision avoidance trajectory, failing to fully consider safety factors such as the road adhesion coefficient and the time-to-collision distance. Furthermore, few studies consider the impact of traffic conditions in adjacent lanes during collision avoidance control. Most simply assume that there are no interfering vehicles in adjacent lanes and use a single critical safety distance as the decision-making metric for whether to execute collision avoidance control, resulting in reduced safety. Therefore, it is also crucial to comprehensively consider the influence of many safety factors for the vehicle's steering and braking coordinated collision avoidance control system. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention proposes an intelligent vehicle steering and braking coordinated collision avoidance control system and method, so as to further improve the vehicle's emergency collision avoidance capability. At the same time, the entire collision avoidance process comprehensively considers the influence of multiple safety factors to ensure the safety of vehicle driving and ride comfort.

[0006] The technical solutions adopted in the present invention are as follows:

[0007] A method for coordinated collision avoidance control of steering and braking of an intelligent vehicle comprises the following steps:

[0008] Step S1: Acquire the real-time vehicle status information and current road information, establish a coordinate system with the vehicle's center of mass as the reference point, and preliminarily plan the vehicle's uniform lane change trajectory based on a quintic polynomial;

[0009] Step S2: Based on the vehicle's uniform lane change trajectory, the lateral acceleration with respect to time expression a is obtained: y (t), the expression for lateral acceleration a y (t) Find the extreme value and set its absolute value |a y-max | as the maximum lateral acceleration threshold;

[0010] Step S3: construct an ideal maximum lateral acceleration model simulating an excellent driver based on the BP neural network, calculate the ideal maximum lateral acceleration that is most comfortable to ride under the current working condition by simulating the ideal maximum lateral acceleration model of an excellent driver, and output its absolute value |a′ y-max |;

[0011] Step S4, based on the absolute value of the ideal maximum lateral acceleration |a′ y-max | and maximum lateral acceleration threshold |a y-max |, after comparison, output the expected maximum lateral acceleration a for lane change collision avoidance in the current road condition y-max-fine :

[0012] When |a′ y-max |≤|a y-max |, let a y-max-fine =a′ y-max ; when |a′ y-max |>|a y-max |, let a y-max-fine =a y-max ;

[0013] Step S5: Consider the current road surface adhesion coefficient μ0 and the tire friction circle constraint conditions that the longitudinal and lateral coupling forces of the vehicle need to satisfy during the emergency collision avoidance process:

[0014]

[0015] Combined with the expected maximum lateral acceleration a y-max-fine, calculate and output the ideal braking deceleration a′ for cooperative collision avoidance x , where g is the acceleration due to gravity;

[0016] Step S6, by introducing the vehicle collision time TTC which can reflect the degree of danger of driving conditions -1 , combined with the driving information and the critical indicator of the vehicle's danger level, the collaborative collision avoidance braking deceleration threshold a is calculated and output by the following formula x-max :

[0017]

[0018] In the above formula, v m is the initial speed of the vehicle, v n is the obstacle speed, S0 is the longitudinal distance between the vehicle and the obstacle, and 0.8 is the vehicle's dangerous level TTC -1 critical indicators;

[0019] Step S7: Based on the ideal braking deceleration a′ x and braking deceleration threshold a x-max , after comparison, the expected braking deceleration a for cooperative collision avoidance under current road conditions is output x-fine : when |a′ x |≤|a x-max |, let a x-fine =a′ x ; when |a′ x |>|a x-max |, let a x-fine =a x-max ;

[0020] Step S8: based on the expected braking deceleration a of the cooperative collision avoidance x-fine , complete the collaborative collision avoidance trajectory planning, expressed as:

[0021]

[0022] In the above formula, x′(t) and y′(t) are the coordinates of the vehicle at time t, v m is the initial speed of the vehicle, t′ m is the completion time of cooperative collision avoidance, X(t′ m ) is the longitudinal displacement of the center of mass of the ego vehicle when the coordinated collision avoidance is completed;

[0023] Step S9: Based on the planned collaborative collision avoidance trajectory, considering that the ego vehicle has completed collision avoidance smoothly and effectively, analyze the critical collision scenario in the collaborative collision avoidance process and calculate the collaborative collision avoidance critical safety distance D. 协同 ;

[0024] Step S10: Based on the critical safety distance D of the collaborative collision avoidance 协同, compare the real-time collected longitudinal distance S0 between the vehicle and the obstacle and the critical safety distance D of the cooperative collision avoidance 协同 The size relationship between the two:

[0025] If S0<D 协同 If a collision is unavoidable, control the vehicle and brake as hard as possible to reduce the damage caused by the collision;

[0026] If S0≥D 协同 , further determine whether the adjacent lane conditions allow, the judgment method is:

[0027] When there is no vehicle passing in the adjacent lane or the speed of the interfering vehicle in front of the adjacent lane is v 邻 With the initial speed v of the vehicle m Satisfy the size relationship: v 邻 ≥v m , then the adjacent lane conditions are considered to be permitted, and the vehicle is controlled to complete the steering and braking coordinated collision avoidance operation; otherwise, further calculation of t′ m Distance value within time D 邻 :

[0028]

[0029] In the above formula, X 邻 (t′ m ) corresponds to t′ m The longitudinal displacement of the interfering vehicle in the adjacent lane during the time (assuming the speed of the vehicle in the adjacent lane remains stable), S 邻 is the longitudinal distance between the vehicle and the interfering vehicle in the adjacent lane, and Δd0 is the safety distance margin.

[0030] By comparing the distance value D 邻 The longitudinal displacement of the vehicle's center of mass X(t′ m ) The size relationship between the two: If D 邻 ≥X(t′ m ), if the traffic conditions in the adjacent lane allow, the vehicle is controlled to complete the coordinated steering and braking to avoid collision; otherwise, it is determined that the collision is unavoidable, and the vehicle is controlled to brake as hard as possible to reduce the collision damage.

[0031] Furthermore, the method for constructing the ideal maximum lateral acceleration model simulating an excellent driver is:

[0032] Step S3.1, experimental data preparation: invite several experienced drivers to drive the test vehicle and conduct multiple collision avoidance tests on different road sections at different speeds and different longitudinal distances between the vehicle and the obstacle. During the test, the real-time status information of the vehicle and the road section information, including the vehicle speed v m , current lane width W0, collision avoidance completion time t m , the corresponding longitudinal displacement of the vehicle X(tm ), the longitudinal distance S0 between the vehicle and the obstacle and the center line of the lane;

[0033] Step S3.2, with [v m , W0, t m ,X(t m ), S0] T and the lane centerline as input, with the absolute value of the ideal maximum lateral acceleration |a′ y-max |As the output, the ideal maximum lateral acceleration model of an excellent driver is trained.

[0034] Furthermore, the method for training the ideal maximum lateral acceleration model of an excellent driver is:

[0035] 1) Initialize the neural network;

[0036] 2) Determine the output of each node in the hidden layer and the output layer, where the output of each node in the hidden layer is:

[0037]

[0038] In the above formula, H j Represents the output of the jth hidden layer, x i represents the input value of the i-th input layer, ω ij represents the weight from the i-th input layer node to the j-th hidden layer node, a j is the bias from the input layer to the hidden layer, n is the number of nodes in the input layer, and g(x) is the activation function;

[0039] The output of each node in the output layer is:

[0040]

[0041] In the above formula, O k represents the output result of the kth output layer, ω jk represents the weight from the jth hidden layer node to the kth output layer node, b k is the bias from the hidden layer to the output layer, l is the number of nodes in the hidden layer;

[0042] 3) Calculate the error of the output result, where the calculation formula of the error E is:

[0043]

[0044] In the above formula, Y k represents the expected output value of the kth output layer, and m is the number of nodes in the output layer;

[0045] 4) Update the weights and biases based on the results calculated in steps 1) to 3);

[0046] 5) Repeat steps 2), 3), and 4) based on the new weights and biases until the error is less than the set error threshold. This means that the algorithm has converged and the training of the ideal maximum lateral acceleration model for an excellent driver is complete.

[0047] Furthermore, the calculation formula for updating the weight is:

[0048]

[0049] The calculation formula for updating the bias is:

[0050]

[0051] In the above formula, η is the learning efficiency.

[0052] Furthermore, the critical safety distance D for cooperative collision avoidance is calculated in step S9. 协同 The specific method is:

[0053] Step S9.1: The critical collision moment between the ego vehicle and the obstacle ahead is t0. The ego vehicle coordinates at this moment are expressed as:

[0054]

[0055] The corresponding vehicle heading angle ψ at this moment m (t0) is expressed as:

[0056]

[0057] Step S9.2: Analyze the critical situation where the vehicle does not collide with the obstacle and calculate the critical safety distance D for cooperative collision avoidance. 协同 :

[0058]

[0059] In the above formula, X(t0) and Y(t0) are the longitudinal displacement and lateral displacement of the center of mass of the vehicle at the critical collision moment, respectively. m is the distance from the center of mass of the vehicle to the front of the vehicle, B m is the vehicle width, B n is the width of the obstacle, Δd0 is the safety distance margin;

[0060] Furthermore, the vehicle's uniform lane change trajectory planned in step S1 is expressed as:

[0061]

[0062] In the above formula, t is the time variable, x(t) and y(t) are the coordinates of the vehicle at time t, and v m is the vehicle speed, W0 is the current lane width, tm is the lane change completion time, X(t m ) is the longitudinal displacement of the vehicle’s center of mass when the lane change is completed;

[0063] In step S2, the second-order derivative of the vehicle's uniform lane change trajectory is obtained to obtain the expression of the lateral acceleration with respect to time:

[0064]

[0065] Then find the extreme value of the above formula and take the absolute value to obtain the maximum lateral acceleration threshold:

[0066]

[0067] An intelligent vehicle steering and braking coordinated collision avoidance control system includes a perception layer, a decision layer, and an execution layer:

[0068] The perception layer is used to obtain real-time status information of the vehicle and surrounding vehicles and current road information, and input the obtained information into the decision layer;

[0069] The decision layer includes a lane change trajectory planning module, an expected maximum lateral acceleration output module, a collaborative collision avoidance expected braking deceleration output module, a collaborative collision avoidance critical safety distance calculation module and a collision avoidance strategy decision module; the lane change trajectory planning module plans a quintic polynomial lane change trajectory based on the information collected by the perception layer, and transmits the planned trajectory to the expected maximum lateral acceleration output module; the expected maximum lateral acceleration output module is used to constrain the lateral acceleration of the vehicle during the collision avoidance process, and output the expected maximum lateral acceleration to the collaborative collision avoidance expected braking deceleration output module; the collaborative collision avoidance expected braking deceleration output module outputs the expected braking deceleration based on the input expected maximum lateral acceleration and the corresponding constraint conditions; the collaborative collision avoidance critical safety distance calculation module calculates the critical safety distance based on the planned collaborative collision avoidance trajectory and the obtained relevant driving information, and transmits the critical safety distance to the collision avoidance strategy decision module; the collision avoidance strategy decision module completes the collision avoidance decision by comprehensively considering the relationship between the current longitudinal distance between the vehicle and the obstacle and the critical safety distance and the traffic conditions of the adjacent lanes;

[0070] The execution layer is used to execute the collision avoidance decision result output by the decision layer.

[0071] Furthermore, the perception layer specifically includes a GPS device, a wheel speed sensor, a millimeter wave radar, a camera, a lidar and a road adhesion coefficient estimator.

[0072] Furthermore, the information collected by the perception layer specifically includes the real-time position information of the vehicle, the speed of the vehicle, the longitudinal distance between the vehicle and the obstacle and the speed of the obstacle, the longitudinal distance between the vehicle and the interfering vehicle in front of the adjacent lane and the speed of the interfering vehicle in front of the adjacent lane, the width of the obstacle, the current lane width and lane centerline, and the current road surface adhesion coefficient.

[0073] Beneficial effects of the present invention:

[0074] 1) In response to the limitations of single collision avoidance control of vehicles, the present invention proposes a collision avoidance strategy based on coordinated control of steering and braking, which is beneficial to improving the vehicle's collision avoidance capability in emergency situations and enhancing the vehicle's collision avoidance safety.

[0075] 2) In the present invention, when deciding the maximum lateral acceleration for vehicle collision avoidance, relevant state variables that have a significant impact on the driver's actual operation process are comprehensively considered, and the ideal maximum lateral acceleration for the most comfortable ride is output through neural network training, making the operation process closer to the driving characteristics of human drivers, and meeting the passengers' comfort requirements as much as possible while ensuring the vehicle's collision avoidance stability.

[0076] 3) In the present invention, when deciding and outputting the expected braking deceleration for collaborative collision avoidance, the influence of two safety factors, the road adhesion coefficient and the vehicle collision distance, is simultaneously considered; when the final collision avoidance is executed, the corresponding decision is made in combination with the critical safety distance for collision avoidance and the traffic conditions of the adjacent lanes, further ensuring the safety of the entire collision avoidance process. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 This is a block diagram of the intelligent vehicle steering and braking coordinated collision avoidance control system of the present invention;

[0078] Figure 2 This is a specific working flow diagram of the module for outputting the expected maximum lateral acceleration and expected braking deceleration when the vehicle avoids collision in the present invention;

[0079] Figure 3 This is a specific workflow diagram of the vehicle collision avoidance strategy decision module and execution layer in the present invention;

[0080] Figure 4 A schematic diagram of a vehicle collision avoidance scenario in the present invention;

[0081] Figure 5 A schematic diagram of a scene when a vehicle and an obstacle collide critically in the present invention;

[0082] Figure 6 Schematic diagram of a scenario where traffic conditions in adjacent lanes allow for this invention. DETAILED DESCRIPTION

[0083] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0084] like Figure 1 As shown, the intelligent vehicle steering and braking coordinated collision avoidance control system proposed in the present invention has an overall structure mainly including a perception layer, a decision layer and an execution layer.

[0085] The perception layer is used to obtain real-time status information of the vehicle and surrounding vehicles as well as current road information. It mainly obtains the real-time position information of the vehicle through GPS equipment, measures the vehicle speed through wheel speed sensors, and obtains the longitudinal distance between the vehicle and the obstacle and the obstacle speed, the longitudinal distance between the vehicle and the interfering vehicle in front of the adjacent lane and the speed of the interfering vehicle in front of the adjacent lane. Cameras and lidars are used to obtain the obstacle width, the current lane width and the lane centerline. The road adhesion coefficient estimator estimates the road adhesion coefficient in real time, and then transmits this real-time collected information to the collision avoidance decision layer through the CAN bus.

[0086] The decision layer is composed of a lane change trajectory planning module, a desired maximum lateral acceleration output module, a cooperative collision avoidance desired braking deceleration output module, a cooperative collision avoidance critical safety distance calculation module, and a collision avoidance strategy decision module. The details are as follows:

[0087] The lane change trajectory planning module receives the real-time vehicle status information and current road information collected by the perception layer to plan a quintic polynomial lane change trajectory and transmits the planned trajectory to the expected maximum lateral acceleration output module. The expected maximum lateral acceleration output module receives the real-time vehicle status information and current road information collected by the perception layer, as well as the trajectory planned by the lane change trajectory planning module, constrains the lateral acceleration of the vehicle during the collision avoidance process, and outputs the expected maximum lateral acceleration to the cooperative collision avoidance expected braking deceleration output module. The cooperative collision avoidance expected braking deceleration output module receives the real-time vehicle status information, current road information, and expected maximum lateral acceleration collected by the perception layer, and outputs the expected braking deceleration based on the input expected maximum lateral acceleration and corresponding constraints. The cooperative collision avoidance critical safety distance calculation module calculates the critical safety distance based on the planned cooperative collision avoidance trajectory and the obtained relevant driving information, and transmits the critical safety distance to the collision avoidance strategy decision module. The collision avoidance strategy decision module makes a collision avoidance decision based on the relationship between the current longitudinal distance between the vehicle and the obstacle and the critical safety distance, as well as the traffic conditions of the adjacent lanes.

[0088] The execution layer is used to execute the collision avoidance decision results output by the decision layer, which mainly includes controlling the vehicle to complete the coordinated steering and braking collision avoidance operations and controlling the vehicle to brake as hard as possible to reduce the collision damage when a collision is unavoidable.

[0089] like Figure 2 As shown, based on the lane change trajectory planning module, the expected maximum lateral acceleration output module, and the collaborative collision avoidance expected braking deceleration output module in the above-mentioned intelligent vehicle steering and braking coordinated collision avoidance control system, the present application proposes an intelligent vehicle steering and braking coordinated collision avoidance control method, which specifically includes the following steps:

[0090] Step S1, such as Figure 4 In the collision avoidance scenario shown, based on the driving and road information obtained by the perception layer, a coordinate system is established with the vehicle's center of mass as the reference point, with the vehicle's forward direction as the x-axis and the direction perpendicular to the vehicle as the y-axis. The uniform lane change trajectory of the vehicle is initially planned based on a quintic polynomial and can be expressed as:

[0091]

[0092] In the above formula, t is the time variable, x(t) and y(t) are the coordinates of the vehicle at time t, and v m is the vehicle speed, W0 is the current lane width, t m is the lane change completion time, X(t m ) is the longitudinal displacement of the vehicle’s center of mass when the lane change is completed;

[0093] Step S2: In the expected maximum lateral acceleration output module, the second-order derivative of the above trajectory expression is obtained to obtain the lateral acceleration expression with respect to time:

[0094]

[0095] Then find the extreme value of the above formula and take the absolute value to obtain the maximum lateral acceleration threshold:

[0096]

[0097] Step S3: Combine the acquired information, use the vehicle's real-time driving information and current road information as input, calculate the ideal maximum lateral acceleration for the most comfortable ride under the current working conditions through the BP neural network-based model of the ideal maximum lateral acceleration of an excellent driver, and output its absolute value |a′ y-max The BP neural network-based model of the ideal maximum lateral acceleration of an excellent driver is trained using a large amount of experimental data on skilled drivers maneuvering vehicles to change lanes and avoid collisions, so that the model reaches a certain predetermined accuracy. The specific steps of the experiment and training are as follows:

[0098] Step S3.1, experimental data preparation: Invite several experienced drivers to drive the test vehicle and conduct multiple collision avoidance tests on different road sections at different speeds and different longitudinal distances between the vehicle and the obstacle. The various information collection devices installed on the test vehicle collect real-time driving information and road information of the road section, including obtaining the real-time vehicle position information through GPS equipment and recording the collision avoidance completion time t m And the corresponding longitudinal displacement X(t m ), wheel speed sensor measures the vehicle speed v m The millimeter-wave radar obtains the longitudinal distance S0 between the vehicle and the obstacle, and the camera and lidar obtain the current lane width W0 and the lane centerline. During the experiment, the computer used to record the data stores and records the experimental data in real time at a frequency of 10Hz.

[0099] Step S3.2, using the above experimental data, [vm ， W0,t m ,X(t m ), S0] T and the lane centerline as input, with the absolute value of the ideal maximum lateral acceleration |a′ y-max | As the output, the ideal maximum lateral acceleration model of an excellent driver is trained. The specific training process is as follows:

[0100] 1) Initialize the neural network, assuming that the number of input layer nodes is n, the number of hidden layer nodes is l, the number of output layer nodes is m, ω ij Represents the weight from the i-th input layer node to the j-th hidden layer node; ω jk Represents the weight from the jth hidden layer node to the kth output layer node; the bias from the input layer to the hidden layer is a j ; The bias from the hidden layer to the output layer is b k ; The learning efficiency is η; the activation function is g(x).

[0101] Among them, the specific activation function is:

[0102] 2) Determine the output of each node in the hidden layer and the output layer, where the output of each node in the hidden layer is:

[0103]

[0104] In the above formula, H j represents the output of the jth hidden layer; x i Represents the input value of the i-th input layer.

[0105] The output of each node in the output layer is:

[0106]

[0107] In the above formula, O k Represents the output result of the kth output layer.

[0108] 3) Calculate the error of the output result, where the calculation formula of the error E is:

[0109]

[0110] In the above formula, Y k Represents the expected output value of the k-th output layer.

[0111] 4) Update the weights and biases based on the results calculated in steps 1) to 3). The formula for updating the weights is:

[0112]

[0113] The calculation formula for updating the bias is:

[0114]

[0115] 5) Repeat steps 2), 3), and 4) based on the new weights and biases until the error is less than the set error threshold. This means that the algorithm has converged and the training of the ideal maximum lateral acceleration model for an excellent driver is complete.

[0116] Step S4, based on the absolute value of the ideal maximum lateral acceleration |a′ y-max | and maximum lateral acceleration threshold |a y-max |, after comparison, output the expected maximum lateral acceleration a for lane change collision avoidance in the current road condition y-max-fine ,Right now:

[0117] When |a′ y-max |≤|a y-max |, let a y-max-fine =a′ y-max ; when |a′ y-max |>|a y-max |, let a y-max-fine =a y-max ;

[0118] Step S5: In the cooperative collision avoidance expected braking deceleration output module, the current road surface adhesion coefficient μ0 and the tire friction circle constraint conditions that the longitudinal and lateral coupling forces of the vehicle need to satisfy during the emergency collision avoidance process are considered:

[0119]

[0120] Combined with the expected maximum lateral acceleration a obtained in the previous step y-max-fine, calculate and output the ideal braking deceleration a′ for cooperative collision avoidance x , where g is the acceleration due to gravity;

[0121] Step S6, by introducing the vehicle collision time TTC which can reflect the degree of danger of driving conditions -1 , combined with the driving information and the critical indicator of the vehicle's danger level, the collaborative collision avoidance braking deceleration threshold a is calculated and output by the following formula x-max :

[0122]

[0123] In the above formula, v m is the initial speed of the vehicle, v n is the obstacle speed, S0 is the longitudinal distance between the vehicle and the obstacle, and 0.8 is the vehicle's dangerous level TTC -1 critical indicators;

[0124] Here, the vehicle collision time TTC introduced -1 , from the definition, when the vehicle reaches a safe state, the relative speed of the two vehicles tends to 0, at this time TTC -1 Approaching 0; the safer the vehicle, the higher the TTC -1 The smaller the value, the higher the vehicle collision risk level, the higher the TTC -1 The larger the value, the greater the -1 The safety level evaluation method is: when TTC -1 <0.5, the vehicle is at a safe level; when 0.5 < TTC -1 <0.8, the vehicle is at the danger warning level; when TTC -1 >0.8, the vehicle is in a dangerous level. Therefore, here we take 0.8 as the vehicle in a dangerous level TTC -1 Critical indicators.

[0125] Step S7: Based on the ideal braking deceleration a′ x and braking deceleration threshold a x-max , after comparison, the expected braking deceleration a for cooperative collision avoidance under current road conditions is output x-fine , that is: when |a′ x |≤|a x-max |, let a x-fine =a′ x ; when |a′ x |>|a x-max |, let a x-fine =a x-max ;

[0126] In step S8, based on the above, a coordinate system is established with the center of mass of the vehicle as the reference point, with the vehicle's forward direction as the x-axis and the vertical direction of the vehicle as the y-axis, to complete the collaborative collision avoidance trajectory planning, which can be specifically expressed as:

[0127]

[0128] In the above formula, x′(t) and y′(t) are the coordinates of the vehicle at time t, v m is the initial speed of the vehicle, t′ m is the completion time of cooperative collision avoidance, X(t′ m ) is the longitudinal displacement of the vehicle’s center of mass when completing the coordinated collision avoidance.

[0129] Step S9, as Figure 5 As shown in Figure 1, based on the planned collaborative collision avoidance trajectory, considering that the ego vehicle smoothly and effectively completes collision avoidance, the critical collision scenario in the collaborative collision avoidance process is analyzed. The specific steps are:

[0130] Step S9.1: The critical collision moment between the ego vehicle and the obstacle ahead is t0. The ego vehicle coordinates at this moment can be expressed as:

[0131]

[0132] The corresponding vehicle heading angle ψ at this moment m (t0) can be expressed as:

[0133]

[0134] Step S9.2: Further, analyze the critical situation where the vehicle does not collide with the obstacle, and calculate the critical safety distance D for cooperative collision avoidance. 协同 :

[0135]

[0136] In the above formula, X(t0) and Y(t0) are the longitudinal displacement and lateral displacement of the center of mass of the vehicle at the critical collision moment, respectively. m is the distance from the center of mass of the vehicle to the front of the vehicle, B m is the vehicle width, B n is the obstacle width, and Δd0 is the safety distance margin.

[0137] Step S10, as Figure 3 As shown, in the collision avoidance strategy decision module and execution layer, based on the critical safety distance D of collaborative collision avoidance 协同 , compare the real-time collected longitudinal distance S0 between the vehicle and the obstacle and the critical safety distance D of the cooperative collision avoidance 协同 The size relationship between the two:

[0138] If S0<D协同 , then the collision is unavoidable, control the vehicle and brake as hard as possible to reduce the collision damage;

[0139] If S0≥D 协同 , then further determine whether the adjacent lane conditions allow it. The specific determination method is:

[0140] Based on the adjacent lane driving information obtained by the millimeter wave radar (assuming that the driving speed of the adjacent lane vehicle remains stable), if there is no vehicle passing through the adjacent lane or the speed of the interfering vehicle in front of the adjacent lane is v 邻 and the vehicle speed v m Satisfy the size relationship: v 邻 ≥v m , then the conditions are obviously allowed, and the vehicles are then controlled to complete the coordinated collision avoidance operation; otherwise, another calculation is performed to obtain t′ m Distance value within time D 邻 ,like Figure 6 As shown:

[0141]

[0142] In the above formula, X 邻 (t′ m ) corresponds to t′ m The longitudinal displacement of the interfering vehicle in front of the adjacent lane during the time, S 邻 is the longitudinal distance between the vehicle and the interfering vehicle in the adjacent lane, and Δd0 is the safety distance margin.

[0143] Then by comparing the distance value D 邻 The longitudinal displacement of the vehicle's center of mass X(t′ m ) The size relationship between the two: If D 邻 ≥X(t′ m ), then the traffic conditions in the adjacent lane allow, and the vehicles are controlled to complete the coordinated collision avoidance maneuver; otherwise, it is determined that the collision is unavoidable, and the vehicles are controlled to brake as hard as possible to reduce the collision damage.

[0144] The above embodiments are intended only to illustrate the design concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. The scope of protection of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design concepts disclosed in the present invention are within the scope of protection of the present invention.

Claims

1. A method for coordinated collision avoidance control of steering and braking of an intelligent vehicle, characterized in that: The steps include: Step S1: Acquire the real-time vehicle status information and current road information, establish a coordinate system with the vehicle's center of mass as the reference point, and preliminarily plan the vehicle's uniform lane change trajectory based on a quintic polynomial; Step S2: Based on the vehicle's uniform lane change trajectory, the lateral acceleration with respect to time expression a is obtained: y (t), the expression for lateral acceleration a y (t) Find the extreme value and set its absolute value |a y-max | as the maximum lateral acceleration threshold; Step S3: construct an ideal maximum lateral acceleration model simulating an excellent driver based on the BP neural network, calculate the ideal maximum lateral acceleration that is most comfortable to ride under the current working condition by simulating the ideal maximum lateral acceleration model of an excellent driver, and output its absolute value |a′ y-max The method for training the ideal maximum lateral acceleration model of an excellent driver is: 1) Initialize the neural network; 2) Determine the output of each node in the hidden layer and the output layer, where the output of each node in the hidden layer is: In the above formula, H j Represents the output of the jth hidden layer, x i represents the input value of the i-th input layer, ω ij represents the weight from the i-th input layer node to the j-th hidden layer node, a j is the bias from the input layer to the hidden layer, n is the number of nodes in the input layer, and g(x) is the activation function; The output of each node in the output layer is: In the above formula, O k Represents the output result of the kth output layer, ω jk represents the weight from the jth hidden layer node to the kth output layer node, b k is the bias from the hidden layer to the output layer, l is the number of nodes in the hidden layer; 3) Calculate the error of the output result, where the calculation formula of the error E is: In the above formula, Y k represents the expected output value of the kth output layer, and m is the number of nodes in the output layer; 4) Update the weights and biases based on the results calculated in steps 1) to 3); 5) Repeat steps 2), 3), and 4) based on the new weights and biases until the error is less than the set error threshold. This means the algorithm has converged and the training of the ideal maximum lateral acceleration model for an excellent driver is complete. The calculation formula for updating the weight is: The calculation formula for updating the bias is: In the above formula, η is the learning efficiency; Step S4, based on the absolute value of the ideal maximum lateral acceleration |a′ y-max | and maximum lateral acceleration threshold |a y-max |, after comparison, output the expected maximum lateral acceleration a for lane change collision avoidance in the current road condition y-max-fine : This |a' y-max |≤|a y-max |Time, Rei a y-max-fine = a' y-max ;This |a' y-max |>|a y-max |Time, Rei a y-max-fine = a y-max ; Step S5: Consider the current road surface adhesion coefficient μ0 and the tire friction circle constraint conditions that the longitudinal and lateral coupling forces of the vehicle need to satisfy during the emergency collision avoidance process: Combined with the expected maximum lateral acceleration a y-max-fine , calculate and output the ideal braking deceleration a′ for cooperative collision avoidance x , where g is the acceleration due to gravity; Step S6, by introducing the vehicle collision time TTC which can reflect the degree of danger of driving conditions -1 , combined with the driving information and the critical indicator of the vehicle's danger level, the collaborative collision avoidance braking deceleration threshold a is calculated and output by the following formula x-max : In the above formula, v m is the initial speed of the vehicle, v n is the obstacle speed, S0 is the longitudinal distance between the vehicle and the obstacle, and 0.8 is the vehicle's dangerous level TTC -1 critical indicators; Step S7: Based on the ideal braking deceleration a′ x and braking deceleration threshold a x-max , after comparison, the expected braking deceleration a for cooperative collision avoidance under current road conditions is output x-fine : when |a′ x |≤|a x-max |, let a x-fine =a′ x ; when |a′ x |>|a x-max |, let a x-fine =a x-max ; Step S8: based on the expected braking deceleration a of the cooperative collision avoidance x-fine , complete the collaborative collision avoidance trajectory planning, expressed as: In the above formula, x′(t) and y′(t) are the coordinates of the vehicle at time t, v m is the initial speed of the vehicle, t′ m is the completion time of cooperative collision avoidance, X(t′ m ) is the longitudinal displacement of the center of mass of the ego vehicle when the coordinated collision avoidance is completed; Step S9: Based on the planned collaborative collision avoidance trajectory, considering that the ego vehicle has completed collision avoidance smoothly and effectively, analyze the critical collision scenario in the collaborative collision avoidance process and calculate the collaborative collision avoidance critical safety distance D. 协同 ; Step S10: Based on the critical safety distance D of the collaborative collision avoidance 协同 , compare the real-time collected longitudinal distance S0 between the vehicle and the obstacle and the critical safety distance D of the cooperative collision avoidance 协同 The size relationship between the two: If S0<D 协同 If a collision is unavoidable, control the vehicle and brake as hard as possible to reduce the damage caused by the collision; If S0≥D 协同 , further determine whether the adjacent lane conditions allow, the judgment method is: When there is no vehicle passing in the adjacent lane or the speed of the interfering vehicle in front of the adjacent lane is v 邻 With the initial speed v of the vehicle m Satisfy the size relationship: v 邻 ≥v m , then the adjacent lane conditions are considered to be permitted, and the vehicle is controlled to complete the steering and braking coordinated collision avoidance operation; otherwise, further calculation of t′ m Distance value within time D 邻 : In the above formula, X 邻 (t′ m ) corresponds to t′ m The longitudinal displacement of the interfering vehicle in front of the adjacent lane during the time, S 邻 is the longitudinal distance between the ego vehicle and the interfering vehicle in front of the adjacent lane, Δd0 is the safety distance margin; By comparing the distance value D 邻 The longitudinal displacement of the vehicle's center of mass X(t′ m ) The size relationship between the two: If D 邻 ≥X(t′ m ), if the traffic conditions in the adjacent lane allow, the vehicle is controlled to complete the coordinated steering and braking to avoid the collision; otherwise, it is determined that the collision is unavoidable, and the vehicle is controlled to brake with all its strength to reduce the collision damage.

2. The method for controlling the coordinated collision avoidance of steering and braking of an intelligent vehicle according to claim 1, characterized in that: The method for constructing the ideal maximum lateral acceleration model simulating an excellent driver is: Step S3.1, experimental data preparation: invite several experienced drivers to drive the test vehicle and conduct multiple collision avoidance tests on different road sections at different speeds and different longitudinal distances between the vehicle and the obstacle. During the test, the real-time status information of the vehicle and the road section information, including the vehicle speed v m , current lane width W0, collision avoidance completion time t m , the corresponding longitudinal displacement of the vehicle X(t m ), the longitudinal distance S0 between the vehicle and the obstacle and the center line of the lane; Step S3.2, with [v m , W0, t m ,X(t m ), S0] T and the lane centerline as input, with the absolute value of the ideal maximum lateral acceleration |a′ y-max |As the output, the ideal maximum lateral acceleration model of an excellent driver is trained.

3. The intelligent vehicle steering and braking coordinated collision avoidance control method according to claim 1, characterized in that: The critical safety distance D for cooperative collision avoidance is calculated in step S9. 协同 The specific method is: Step S9.1: The critical collision moment between the ego vehicle and the obstacle ahead is t0. The ego vehicle coordinates at this moment are expressed as: The corresponding vehicle heading angle ψ at this moment m (t0) is expressed as: Step S9.2: Analyze the critical situation where the vehicle does not collide with the obstacle and calculate the critical safety distance D for cooperative collision avoidance. 协同 : In the above formula, X(t0) and Y(t0) are the longitudinal displacement and lateral displacement of the center of mass of the vehicle at the critical collision moment, respectively. m is the distance from the center of mass of the vehicle to the front of the vehicle, B m is the vehicle width, B n is the obstacle width, and Δd0 is the safety distance margin.

4. The intelligent vehicle steering and braking coordinated collision avoidance control method according to claim 1, characterized in that: The vehicle's uniform lane change trajectory planned in step S1 is expressed as: In the above formula, t is the time variable, x(t) and y(t) are the coordinates of the vehicle at time t, and v m is the vehicle speed, W0 is the current lane width, t m is the lane change completion time, X(t m ) is the longitudinal displacement of the vehicle’s center of mass when the lane change is completed; In step S2, the second-order derivative of the vehicle's uniform lane change trajectory is obtained to obtain the expression of the lateral acceleration with respect to time: Then find the extreme value of the above formula and take the absolute value to obtain the maximum lateral acceleration threshold: 。 5. An intelligent vehicle steering and braking coordinated collision avoidance control system based on the intelligent vehicle steering and braking coordinated collision avoidance control method according to claim 1, characterized in that: Includes perception layer, decision layer and execution layer; The perception layer is used to obtain real-time status information of the vehicle and surrounding vehicles and current road information, and input the obtained information into the decision layer; The decision layer includes a lane change trajectory planning module, an expected maximum lateral acceleration output module, a collaborative collision avoidance expected braking deceleration output module, a collaborative collision avoidance critical safety distance calculation module and a collision avoidance strategy decision module; the lane change trajectory planning module plans a quintic polynomial lane change trajectory based on the information collected by the perception layer, and transmits the planned trajectory to the expected maximum lateral acceleration output module; the expected maximum lateral acceleration output module is used to constrain the lateral acceleration of the vehicle during the collision avoidance process, and output the expected maximum lateral acceleration to the collaborative collision avoidance expected braking deceleration output module; the collaborative collision avoidance expected braking deceleration output module outputs the expected braking deceleration based on the input expected maximum lateral acceleration and the corresponding constraint conditions; the collaborative collision avoidance critical safety distance calculation module calculates the critical safety distance based on the planned collaborative collision avoidance trajectory and the obtained relevant driving information, and transmits the critical safety distance to the collision avoidance strategy decision module; the collision avoidance strategy decision module completes the collision avoidance decision by comprehensively considering the relationship between the current longitudinal distance between the vehicle and the obstacle and the critical safety distance and the traffic conditions of the adjacent lanes; The execution layer is used to execute the collision avoidance decision result output by the decision layer.

6. The intelligent vehicle steering and braking coordinated collision avoidance control system according to claim 5, characterized in that: The perception layer specifically includes GPS equipment, wheel speed sensors, millimeter wave radar, cameras, lidar and road adhesion coefficient estimator.

7. The intelligent vehicle steering and braking coordinated collision avoidance control system according to claim 6, characterized in that: The information collected by the perception layer specifically includes the real-time position information of the vehicle, the speed of the vehicle, the longitudinal distance between the vehicle and the obstacle and the speed of the obstacle, the longitudinal distance between the vehicle and the interfering vehicle in front of the adjacent lane and the speed of the interfering vehicle in front of the adjacent lane, the width of the obstacle, the current lane width and lane centerline, and the current road surface adhesion coefficient.

Images