Obstacle avoidance method for unmanned vehicle platooning based on sliding mode control and evaluation index

Abstract

The present invention discloses an unmanned vehicle formation obstacle avoidance method based on sliding mode control and evaluation index. First, an unmanned vehicle pilot-follower formation model based on distance-angle is established, and a fixed-time sliding mode controller is designed so that the trajectory tracking error of the system can converge to zero along the proposed fixed-time sliding mode surface, and does not depend on the initial value of the system. The system has good stability and convergence, and the formation generation and maintenance capabilities of the unmanned vehicle formation are effectively improved. Secondly, the overall formation obstacle avoidance method is adopted, and an unmanned vehicle formation dynamic transformation obstacle avoidance strategy considering environmental constraints is proposed. A formation knowledge base is established, three formation obstacle avoidance modes of zero formation transformation, formation contraction transformation and formation switching transformation are designed, formation transformation performance evaluation indicators are designed, and a formation formation dynamic transformation decision-making mechanism is established, so that the unmanned vehicle formation can make the optimal formation transformation selection under different obstacle environments, effectively improving the obstacle avoidance efficiency of the unmanned vehicle formation.

Description

Technical field:

[0001] The patent of this invention relates to the field of obstacle avoidance control for unmanned vehicle formations, and more specifically, to an obstacle avoidance method for unmanned vehicle formations based on sliding mode control and evaluation indicators. Background technology:

[0002] With the development of artificial intelligence and autonomous driving technology, unmanned vehicle formations have broad application prospects in industrial production, disaster search and rescue, park inspection and other fields. When unmanned vehicle formations perform specific tasks in an obstacle environment, they need to maintain the desired formation movement and avoid obstacles reasonably and efficiently. This requires the unmanned vehicle formation to have stable formation formation generation and formation maintenance capabilities, and to make the best obstacle avoidance strategy according to the actual environment. At present, the commonly used formation control methods include behavior-based method, virtual structure method, pilot following method, etc. The pilot following method has a simple control structure, easy formation maintenance, and strong scalability. It is a commonly used formation control method. Faced with special environments such as disaster search and rescue where the road conditions are complex and the signal transmission is not smooth, errors are inevitable in the communication between the pilot vehicle and the follower vehicle. At this time, in order to smoothly carry out the search and rescue mission and achieve high-efficiency and high-precision formation control, it is necessary to ensure continuous and stable communication between the follower vehicle and the pilot vehicle. In response to this demand, it is particularly important to calculate the posture information between the unmanned vehicles and provide formation control evaluation indicators. Common obstacle avoidance methods include single obstacle avoidance, split obstacle avoidance, and overall obstacle avoidance. The overall obstacle avoidance method will not disrupt the formation in the process of avoiding obstacles, which can meet the dual requirements of formation obstacle avoidance and maintaining formation, and the integrity and stability of the formation can be effectively guaranteed.

[0003] Patent No. CN113328663A discloses a dual closed-loop sliding mode control method for a permanent magnet synchronous motor based on parameter optimization. This method uses a genetic algorithm to optimize the initial parameters of the sliding mode control, and outputs a PWM signal through vector modulation to control the running speed of the vehicle. This method improves the reliability of the sliding film controller parameter selection, but manual debugging takes too long and the control performance is not high.

[0004] Patent No. CN108919837A discloses a second-order sliding mode control method for autonomous driving vehicles based on visual dynamics. This method establishes a two-degree-of-freedom vehicle lateral dynamics model and a visual dynamics model. On the basis of the first-order sliding mode controller, by reconstructing the sliding mode variables, the value range of the weighted coefficient in the control law of the second-order sliding mode controller is obtained, and the weighted coefficient is applied to the control law to realize autonomous driving. This method improves the robustness of the vehicle, but it is impossible to achieve high-precision obstacle avoidance by relying on a single vision, and the actual application effect is not good.

[0005] Patent No. CN111650929A discloses an adaptive sliding mode control method, system and mobile robot controller. This method combines the advantages of adaptive control and sliding mode control, and designs a sliding mode control law with adaptive adjustment of switching gain. This method can make the mobile robot converge to the desired trajectory and desired speed quickly, improving the efficiency of vehicle driving. However, adaptive control cannot guarantee that the system remains stable during the parameter adjustment process, and there are certain limitations.

[0006] Patent No. CN114371711A discloses a robot formation obstacle avoidance path planning method. The invention uses a grid method to construct a space for robot formation path planning, and uses the VSF-A* algorithm's variable formation path planning algorithm to search in the constructed space to obtain a feasible formation path. Through this method, the robot can successfully avoid obstacles. However, this method destroys the integrity of the robot formation, which weakens the adaptability of the robot formation.

[0007] Patent No. CN112987732A discloses a multi-mobile robot formation obstacle avoidance control method based on the artificial potential field method. This method establishes a formation database, obtains the path and width available for passage in the obstacle area through the potential field function, and selects a suitable formation from the formation database for obstacle avoidance. This method improves the stability of the formation obstacle avoidance system and the overall efficiency of the formation obstacle avoidance, but the artificial potential field method has local minima and unreachable target problems, and has certain application limitations. Summary of the invention:

[0008] In view of the problems of poor formation accuracy, difficulty in maintaining formation, loss of overall obstacle avoidance, and insufficient stability in the current unmanned vehicle formation, the present invention proposes an unmanned vehicle formation obstacle avoidance control method based on fixed-time sliding mode control. In view of the problems of poor formation generation and maintenance accuracy and difficulty in maintaining formation of unmanned vehicle formation, an unmanned vehicle leading-following formation model based on distance-angle is established, and a fixed-time sliding mode controller is designed so that the trajectory tracking error of the system can converge to zero along the proposed fixed-time sliding mode surface and does not depend on the initial value of the system. The system has good stability and convergence, and the formation generation and maintenance capabilities of the unmanned vehicle formation are effectively improved. In view of the problems of loss of overall obstacle avoidance and easy collision of unmanned vehicle formations in obstacle environments, the overall obstacle avoidance method of the formation is adopted, and an unmanned vehicle formation dynamic transformation obstacle avoidance strategy considering environmental constraints is proposed. Establish a formation knowledge base, design three formation obstacle avoidance modes: zero formation transformation, formation contraction transformation and formation switching transformation, design formation transformation performance evaluation indicators, and establish a formation dynamic transformation decision-making mechanism, so that the unmanned vehicle formation can make the optimal formation transformation choice in different obstacle environments, effectively improving the obstacle avoidance efficiency of the unmanned vehicle formation.

[0009] The purpose of the present invention is achieved through the following technical solutions: a method for unmanned vehicle formation obstacle avoidance based on sliding mode control and evaluation indicators, the method mainly includes three parts: unmanned vehicle formation control algorithm design, unmanned vehicle formation overall obstacle avoidance strategy design and formation recovery.

[0010] 1. Design of unmanned vehicle formation control algorithm:

[0011] 1. Establish a pilot-follower formation model. The pilot unmanned vehicle i moves in a two-dimensional plane according to the given linear velocity v and the given angular velocity ω, and the following unmanned vehicle j only needs to maintain the desired relative distance and relative angle The desired formation can be formed by following the leading unmanned vehicle. Through this control structure, the entire unmanned vehicle formation control problem can be transformed into the trajectory tracking problem of the leading unmanned vehicle by the following unmanned vehicle.

[0012] The kinematic model of each unmanned vehicle is shown in Equation 1:

[0013]

[0014] Subscript j represents the jth following unmanned vehicle, L r is the distance between the UGV center of mass and the wheelbase, q = (x, y, θ) T Represents the pose vector of the unmanned vehicle in the (x, O, y) coordinate system. Indicates that the unmanned vehicle is at the reference trajectory point The pose vector of L. ij represents the relative distance between the following unmanned vehicle and the leading unmanned vehicle, ψ ij It represents the relative turning angle between the following unmanned vehicle and the leading unmanned vehicle. θ represents the direction of the unmanned vehicle's movement, that is, the angle between the velocity vector v and the positive direction of the θ axis, which is also the heading angle of the unmanned vehicle.

[0015] Through the operation of spatial geometric relations, the trajectory tracking error model of the unmanned vehicle formation can be obtained as follows:

[0016]

[0017]

[0018] in, represents the lateral trajectory tracking error of the unmanned vehicle formation, represents the longitudinal trajectory tracking error of the unmanned vehicle formation, Represents the trajectory tracking error of the unmanned vehicle formation.

[0019]

[0020] 2. Design a fixed-time sliding mode controller. The unmanned vehicle pilot-follower position error established in the previous step effectively transforms the unmanned vehicle formation control problem into a trajectory tracking control problem of the following unmanned vehicle to the pilot unmanned vehicle. Next, a fixed-time sliding mode controller needs to be designed so that the trajectory tracking error between the following unmanned vehicle and the pilot unmanned vehicle is Converges to the smallest neighborhood of the origin.

[0021] The designed sliding surface and fixed-time sliding mode control law are:

[0022]

[0023]

[0024] After a series of mathematical operations, the final fixed-time sliding mode control law can be obtained as:

[0025]

[0026] in, α and β are fixed parameters of the fixed-time sliding mode control law.

[0027] The designed fixed-time sliding mode controller enables the unmanned vehicle formation to track the trajectory of the leading unmanned vehicle within a fixed time, and the trajectory tracking error of each follower unmanned vehicle can converge to zero quickly within a fixed time and does not depend on the initial value of the system. The fixed-time sliding mode controller has excellent control performance. Each unmanned vehicle can form a specified desired formation from any initial state and maintain the formation effect. The formation generation and maintenance capabilities of the unmanned vehicle formation are effectively improved, and the formation control performance is significantly improved.

[0028] 3. Establish a formation knowledge base. The purpose of establishing a formation knowledge base is to enable the unmanned vehicle formation to directly select the appropriate formation from the established formation knowledge base for obstacle avoidance according to the formation transformation mode during the obstacle avoidance process. First, each unmanned vehicle is assigned a unique number to mark different unmanned vehicles, such as Q1 and Q2. The present invention uses a pilot-follower formation model, and the pilot unmanned vehicle is numbered 1, and the other following unmanned vehicles are numbered 2, 3..., n, where n is the number of vehicles in the unmanned vehicle formation. A parameter matrix is ​​established to represent the positional relationship between the unmanned vehicles, and the expression relationship is shown in the following formula:

[0029]

[0030] Among them, the matrix H represents the parameter value of a formation in the unmanned vehicle formation knowledge base. sj Indicates the jth unmanned vehicle Q in this formation j The state value of . In the matrix H sj In, h 1j represents the j-th unmanned vehicle Qj Number information, h 2j represents j unmanned vehicles Q j The number of the pilot unmanned vehicle Q1 to be tracked, h 3j Represents the driverless car Q j The expected relative distance from the pilot unmanned vehicle Q1, h 4j Indicates Q j The expected relative angle with Q1. By establishing a formation knowledge base, the unmanned vehicle formation can quickly and effectively switch formations when switching formations to avoid obstacles, thus passing through the obstacle area.

[0031] 2. Design of overall obstacle avoidance strategy for unmanned vehicle formation

[0032] The overall obstacle avoidance of the formation mainly includes two methods: formation contraction obstacle avoidance and formation switching obstacle avoidance. Formation contraction obstacle avoidance means that after the unmanned vehicle formation encounters an obstacle, it calculates the size of the passable obstacle area and its own formation width. If the length of the passable obstacle area is less than the length of its own formation and the width of the formation is compressed, it can pass through the obstacle area without collision. At this time, the unmanned vehicle formation continues to maintain the original formation topology structure, that is, the formation shape remains unchanged. While ensuring that there is no collision between unmanned vehicles, the original expected formation is reasonably compressed, and then the obstacle area is passed. After passing the obstacle area, the formation formation is restored to its original state. Formation switching obstacle avoidance means that when the unmanned vehicle formation cannot pass through the obstacle area by formation contraction obstacle avoidance transformation, according to the formation transformation strategy, the formation topology structure is changed, that is, the original formation shape is changed to pass through the obstacle area, and then the formation is restored to the original expected formation. The steps for the two formation transformations are as follows:

[0033] 1. Establish evaluation indicators for formation contraction and obstacle avoidance

[0034] (1) Establish the concept of the deformation scale of the unmanned vehicle external formation to describe the degree to which the width of the current unmanned vehicle formation needs to be compressed.

[0035]

[0036] Among them, λ represents the deformation scale of the unmanned vehicle formation, d o Represents the actual passable environment width of the current obstacle area, and L represents the width of the current unmanned vehicle formation.

[0037] (2) Establish the concept of compression scale to describe the allowable degree of formation expansion and contraction

[0038]

[0039] Where ζ represents the compression scale, δ represents the safe distance between unmanned vehicles, and σ represents the minimum distance between two adjacent unmanned vehicles in the current formation.

[0040] 2. Establishing evaluation indicators for formation switching and obstacle avoidance

[0041] (1) Establish a formation transformation rate index to describe the positional geometric change between the transformed formation and the expected formation before the transformation in the process of the unmanned vehicle formation avoiding obstacles. M d express:

[0042]

[0043] Among them, X r represents the formation matrix after the formation transformation, X p represents the expected formation of the unmanned vehicle platoon before the transformation, L Δ =X p -X r , represents the change in the geometric relationship of the formation matrix.

[0044] (2) Establish a formation convergence time ratio index to describe how much formation change time accounts for the total time of the unmanned vehicle formation's overall obstacle avoidance process, using D u express:

[0045]

[0046] Where T r represents the formation change time of the unmanned vehicle formation after encountering an obstacle, T p It represents the total time of the entire obstacle avoidance process from the formation change to the restoration of the original formation after the obstacle avoidance is completed.

[0047] 3. Establish a dynamic change decision-making mechanism for the overall obstacle avoidance of the formation

[0048] (1) The formation does not change

[0049] When λ≥1, it indicates that the maximum passable distance in the obstacle area is larger than the current width of the unmanned vehicle formation. At this time, the unmanned vehicle formation does not need to change its formation and can pass through the obstacle area in the original formation, and the stability of the formation is guaranteed.

[0050] (2) Formation contraction to avoid obstacles

[0051] When ζ≤λ≤1, it means that the maximum travel distance in the current obstacle area is less than the width of the unmanned vehicle formation and the degree to which the formation needs to be compressed is less than the maximum degree to which it can be compressed, indicating that the obstacle area can be passed by the formation contraction method. The unmanned vehicle formation needs to contract the formation width to λ times the current formation width to pass the obstacle area. Compared with the formation switching obstacle avoidance, the formation contraction obstacle avoidance method can ensure the stability and integrity of the formation.

[0052] (3) Formation switching to avoid obstacles

[0053] When λ≤ζ, it means that the maximum travel distance in the current obstacle area is less than the current formation width and the degree to which the formation needs to be compressed is greater than the degree to which the formation can be compressed. If the formation is compressed to avoid obstacles, collisions will occur between unmanned vehicles. At this time, the formation switching obstacle avoidance method should be selected to pass through the obstacle area. Assuming that there are multiple formations with different widths in the formation knowledge base, first select the switching formations that meet the conditions, calculate the optimal formation through various evaluation indicators, and then perform formation switching to avoid obstacles.

[0054] 3. The formation is restored.

[0055] After the leading unmanned vehicle detects that the last unmanned vehicle has passed the obstacle area, it issues a formation recovery command. The unmanned vehicle readjusts its own position and returns to the expected formation before the formation change to continue the mission.

[0056] The beneficial effects of the present invention are:

[0057] Aiming at the problems of poor formation accuracy and formation maintenance difficulty in the formation generation and maintenance of unmanned vehicle formations, the present invention establishes an unmanned vehicle pilot-follower formation model based on distance-angle, and converts the formation control problem into an unmanned vehicle trajectory tracking problem. A fixed-time sliding mode controller is designed so that the trajectory tracking error of the system can converge to zero along the proposed fixed-time sliding mode surface, and does not depend on the initial value of the system. The system has good stability and timing convergence. The unmanned vehicle can form a specified desired formation from any initial state and maintain the formation effect, and the formation control performance is effectively improved. Aiming at the problems of loss of integrity of the unmanned vehicle formation obstacle avoidance formation and easy collision in obstacle environment, the overall obstacle avoidance mode of the formation is adopted, and the overall obstacle avoidance strategy of the unmanned vehicle formation considering environmental constraints is designed. A formation knowledge base is established, three formation obstacle avoidance modes of zero formation transformation, formation contraction transformation and formation switching transformation are designed, the formation transformation performance evaluation index is set, and a formation dynamic transformation decision mechanism is established, so that the unmanned vehicle formation can make the optimal formation transformation selection under different obstacle environments, and the integrity of the formation is effectively guaranteed. From an economic perspective, this method effectively reduces the risk of collision when unmanned vehicles pass through obstacle areas and improves the obstacle avoidance efficiency of unmanned vehicle formations. Description of the drawings:

[0058] Figure 1 This is a model diagram of the leader-follower formation;

[0059] Figure 2 To optimize the leader-follower formation model schematic;

[0060] Figure 3 A simplified schematic diagram of the time-varying expected relative angle;

[0061] Figure 4 This is a flow chart of the unmanned vehicle formation control method;

[0062] Figure 5 This is the flow chart of the overall obstacle avoidance strategy for the unmanned vehicle formation. Specific implementation method:

[0063] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The elements and features described in one embodiment of the present invention may be combined with the elements and features shown in one or more other embodiments. It should be noted that for the purpose of clarity, the representation and description of components and processes that are not related to the present invention and are known to those of ordinary skill in the art are omitted in the description. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0064] The invention provides an obstacle avoidance control method for unmanned vehicle formation based on a fixed-time sliding mode controller. The method mainly includes three parts: unmanned vehicle formation control algorithm design, unmanned vehicle formation overall obstacle avoidance strategy design and formation recovery.

[0065] Firstly, the present invention aims at solving the problems of poor formation accuracy and difficulty in formation maintenance during the formation generation, maintenance and transformation of unmanned vehicle formations. The unmanned vehicle formation problem is converted into a trajectory tracking control problem of following the unmanned vehicles, an unmanned vehicle pilot-follower formation model is established, and a fixed-time sliding mode formation controller is designed. The fixed-time sliding mode controller can converge the tracking error of the unmanned vehicle formation to zero within a fixed time, so that the system has good stability and convergence, and the formation generation and maintenance capabilities of the unmanned vehicle formation are significantly improved.

[0066] Secondly, the present invention aims at the problem that the formation of unmanned vehicles is destroyed and the integrity of the formation is lost during the obstacle avoidance process. The overall obstacle avoidance method of the formation is adopted, the overall obstacle avoidance strategy of the formation is designed, the formation knowledge base is established, the overall obstacle avoidance performance evaluation index of the formation is designed, and the dynamic transformation decision mechanism of the formation is established. According to different obstacle environments, the optimal formation transformation decision is made. The unmanned vehicle formation can effectively avoid obstacle areas, and the formation and integrity of the formation are effectively guaranteed.

[0067] Finally, the formation is restored. After the leading unmanned vehicle detects that the last unmanned vehicle has passed the obstacle area, it issues a formation restoration command, and the unmanned vehicle readjusts its own position and restores to the desired formation before the formation change, and continues to perform the task.

[0068] 1. The specific scheme for designing the control algorithm of unmanned vehicle formation is as follows:

[0069] 1. Establish a pilot-follower formation model. The pilot unmanned vehicle i moves in a two-dimensional plane according to the given linear velocity v and the given angular velocity ω, and the following unmanned vehicle j only needs to maintain the desired relative distance and relative angle The desired formation can be formed by following the pilot unmanned vehicle. Through this control structure, the entire unmanned vehicle formation control problem can be transformed into the trajectory tracking problem of the pilot unmanned vehicle by the following unmanned vehicle. Figure 1 shown.

[0070] The kinematic model of each unmanned vehicle is shown in Equation 1:

[0071]

[0072] Subscript j represents the jth following unmanned vehicle, L r is the distance between the UGV center of mass and the wheelbase, q = (x, y, θ) T Represents the pose vector of the unmanned vehicle in the (x, O, y) coordinate system. Indicates that the unmanned vehicle is at the reference trajectory point The pose vector of L. ij represents the relative distance between the following unmanned vehicle and the leading unmanned vehicle, ψ ij It represents the relative turning angle between the following unmanned vehicle and the leading unmanned vehicle. θ represents the direction of the unmanned vehicle's movement, that is, the angle between the velocity vector v and the positive direction of the θ axis, which is also the heading angle of the unmanned vehicle.

[0073] Through the operation of spatial geometric relations, the trajectory tracking error model of the unmanned vehicle formation can be obtained as follows:

[0074]

[0075]

[0076] in, represents the lateral trajectory tracking error of the unmanned vehicle formation, represents the longitudinal trajectory tracking error of the unmanned vehicle formation, Represents the trajectory tracking error of the unmanned vehicle formation.

[0077]

[0078] The traditional pilot-following method sends the position information of the pilot vehicle to the follower unmanned vehicle, which has the problem of error transmission. Based on the traditional pilot-following method, the present invention transmits the tracking trajectory of the pilot unmanned vehicle to all the follower unmanned vehicles. When the pilot unmanned vehicle is disturbed and the motion state changes greatly, the follower unmanned vehicle will continue to drive according to the tracking trajectory sent by the pilot unmanned vehicle. This algorithm first determines the main navigator of the formation and the formation formation of the multi-intelligent system. The navigator clarifies the target point or formation path trajectory of the entire formation system, and the follower clarifies its own navigator information. Then, the navigator calculates the expected position of the follower according to its own position and formation formation, and publishes it to the remaining followers through the communication protocol. At the same time, the navigator obtains the status information of the remaining followers, and the navigator adjusts its own speed according to the status information of the followers to ensure that it is within the communication range of the followers. In the process of the navigator driving to the target point, the obstacle avoidance algorithm is used to perform path planning to avoid obstacles. The follower clarifies its own position information and obtains the information of its navigator and the position information of the expected follower. The follower drives to the expected follower position through the obstacle avoidance algorithm for path planning. The formation diagram is as follows Figure 2 shown.

[0079] On this basis, the concepts of time-varying relative distance and time-varying relative angle are introduced to optimize the accuracy of the following unmanned vehicle when following the leading unmanned vehicle.

[0080] 1.1 Time-varying relative distance

[0081] The expected relative distance refers to the minimum distance that two unmanned vehicles should maintain while ensuring driving safety. In the actual driving process of unmanned vehicles, the expected relative distance between the two unmanned vehicles is time-varying. The expected relative distance is related to the relative speed of the leading vehicle and the following vehicle, the speed of the leading vehicle, and the speed of the following vehicle, as shown in Equation 5.

[0082]

[0083] Where S is the desired distance between the two vehicles (m); vrel is the relative speed of the two vehicles (m / s), v f Represents the speed difference between the leading vehicle and the following unmanned vehicle.

[0084] For the following control problem of the following vehicle in this article, the speed of the leading vehicle is given, and the speed of the following vehicle is obtained by the control system. Considering that one of the purposes of designing the controller is to make the speed of the following vehicle infinitely close to the leading vehicle, Formula 5 can be simplified, that is, vrel is set to 0, and Formula 6 can be obtained.

[0085] L d =(0.851×vl+1.6) (6)

[0086] Where L dis the expected relative distance between the two unmanned vehicles, and vl is the speed of the pilot vehicle. Because the unit of speed in the simulation is m / s, but on the actual highway, the unit of car speed is km / h, a unit conversion is performed when deriving the expected relative distance.

[0087] 1.2 Time-varying relative angle

[0088] The unmanned vehicle does not travel in a straight line on the actual road, but its trajectory may be an arbitrary curve. When the pilot vehicle changes direction, the relative angle between the two unmanned vehicles also changes. Figure 3 It is a simplified schematic diagram of the time-varying expected relative angle. Assume that the rectangle represents the pilot car, and the pilot car travels from position a to position . Assume that the ellipse represents the follower car, and the follower car travels from position c to position d. The expected relative angle of the two unmanned vehicles when they are driving is defined as when the two unmanned vehicles just start to drive. The initial expected relative angle is given in the simulation, and the difference between the two is expressed as Equation 7.

[0089]

[0090] in, and are the initial expected relative angles of the two unmanned vehicles. Assuming the bending radius is r, according to Figure 1 We can get formula 8.

[0091]

[0092] According to the position relationship between the two vehicles, Figure 3 As shown, we can get formula 9.

[0093]

[0094] Substituting equations 7 and 8 into equation 9, we can get the desired relative angle As shown in formula 10.

[0095]

[0096] It can be seen from formula 10 that the expected relative angle is related to the expected relative distance, the speed of the pilot vehicle and the angular velocity, and it is a time-varying variable.

[0097] 2. Design a fixed-time sliding mode controller. The unmanned vehicle pilot-follower position error established in the previous step effectively transforms the unmanned vehicle formation control problem into a trajectory tracking control problem of the following unmanned vehicle to the pilot unmanned vehicle. Next, a fixed-time sliding mode controller needs to be designed so that the trajectory tracking error between the following unmanned vehicle and the pilot unmanned vehicle is Converges to the smallest neighborhood of the origin.

[0098] The designed sliding surface and fixed-time sliding mode control law are:

[0099]

[0100]

[0101] After a series of mathematical operations, the final fixed-time sliding mode control law can be obtained as:

[0102]

[0103] in, α and β are fixed parameters of the fixed-time sliding mode control law.

[0104] The designed fixed-time sliding mode controller enables the unmanned vehicle formation to track the trajectory of the leading unmanned vehicle within a fixed time, and the trajectory tracking error of each follower unmanned vehicle can converge to zero quickly within a fixed time and does not depend on the initial value of the system. The fixed-time sliding mode controller has excellent control performance. Each unmanned vehicle can form a specified desired formation from any initial state and maintain the formation effect. The formation generation and maintenance capabilities of the unmanned vehicle formation are effectively improved, and the formation control performance is significantly improved.

[0105] 3. Establish a formation knowledge base. The purpose of establishing a formation knowledge base is to enable the unmanned vehicle formation to directly select the appropriate formation from the established formation knowledge base for obstacle avoidance according to the formation transformation mode during the obstacle avoidance process. First, each unmanned vehicle is assigned a unique number to mark different unmanned vehicles, such as Q1 and Q2. The present invention uses a pilot-follower formation model, and the pilot unmanned vehicle is numbered 1, and the other following unmanned vehicles are numbered 2, 3..., n, where n is the number of vehicles in the unmanned vehicle formation. A parameter matrix is ​​established to represent the positional relationship between the unmanned vehicles, and the expression relationship is shown in the following formula:

[0106]

[0107] Among them, the matrix H represents the parameter value of a formation in the unmanned vehicle formation knowledge base. sj Indicates the jth unmanned vehicle Q in this formation j The state value of . In the matrix H sj In, h 1j represents the j-th unmanned vehicle Q j Number information, h 2j represents j unmanned vehicles Q j The number of the pilot unmanned vehicle Q1 to be tracked, h 3j Represents the driverless car Q j The expected relative distance from the pilot unmanned vehicle Q1, h 4j Indicates Q jThe expected relative angle with Q1. By establishing a formation knowledge base, the unmanned vehicle formation can quickly and effectively switch formations when switching formations to avoid obstacles, thus passing through the obstacle area.

[0108] like Figure 4 As shown, the process of the unmanned vehicle formation control method is as follows:

[0109] First, the unmanned vehicle formation pilot-follower posture error model is established, and relevant information is given, including the initial linear velocity and angular velocity of the pilot unmanned vehicle, and the expected relative distance and relative angle between the following unmanned vehicle and the pilot unmanned vehicle. The pilot unmanned vehicle sends its own posture information to the following unmanned vehicle. The following unmanned vehicle calculates the relative distance and relative angle between itself and the pilot unmanned vehicle based on the received information, and determines whether the relative distance and relative angle at this time are the expected values. If so, it maintains the current motion state to follow the pilot unmanned vehicle. If the expected formation cannot be achieved, the fixed-time sliding mode controller calculates the current posture error information between the pilot unmanned vehicle and the following unmanned vehicle, and outputs the new linear velocity and angular velocity, so that the following unmanned vehicle follows the pilot unmanned vehicle at the set expected relative position and relative angle.

[0110] 2. The specific plan for the overall obstacle avoidance strategy design of the unmanned vehicle formation is:

[0111] The overall obstacle avoidance of the formation mainly includes two methods: formation contraction obstacle avoidance and formation switching obstacle avoidance. Formation contraction obstacle avoidance means that after the unmanned vehicle formation encounters an obstacle, it calculates the size of the passable obstacle area and its own formation width. If the length of the passable obstacle area is less than the length of its own formation and the width of the formation is compressed, it can pass through the obstacle area without collision. At this time, the unmanned vehicle formation continues to maintain the original formation topology structure, that is, the formation shape remains unchanged. While ensuring that there is no collision between unmanned vehicles, the original expected formation is reasonably compressed, and then the obstacle area is passed. After passing the obstacle area, the formation formation is restored to its original state. Formation switching obstacle avoidance means that when the unmanned vehicle formation cannot pass through the obstacle area by formation contraction obstacle avoidance transformation, according to the formation transformation strategy, the formation topology structure is changed, that is, the original formation shape is changed to pass through the obstacle area, and then the formation is restored to the original expected formation. The steps for the two formation transformations are as follows:

[0112] 1. Establish evaluation indicators for formation contraction and obstacle avoidance

[0113] (1) Establish the concept of the deformation scale of the unmanned vehicle external formation to describe the degree to which the width of the current unmanned vehicle formation needs to be compressed.

[0114]

[0115] Among them, λ represents the deformation scale of the unmanned vehicle formation, d oRepresents the actual passable environment width of the current obstacle area, and L represents the width of the current unmanned vehicle formation.

[0116] (2) Establish the concept of compression scale to describe the allowable degree of formation expansion and contraction

[0117]

[0118] Where ζ represents the compression scale, δ represents the safe distance between unmanned vehicles, and σ represents the minimum distance between two adjacent unmanned vehicles in the current formation.

[0119] 2. Establishing evaluation indicators for formation switching and obstacle avoidance

[0120] (1) Establish a formation transformation rate index to describe the positional geometric change between the transformed formation and the expected formation before the transformation in the process of the unmanned vehicle formation avoiding obstacles. M d express:

[0121]

[0122] Among them, X r represents the formation matrix after the formation transformation, X p represents the expected formation of the unmanned vehicle platoon before the transformation, L Δ =X p -X r , represents the change in the geometric relationship of the formation matrix.

[0123] (2) Establish a formation convergence time ratio index to describe how much formation change time accounts for the total time of the unmanned vehicle formation's overall obstacle avoidance process, using D u express:

[0124]

[0125] Where T r represents the formation change time of the unmanned vehicle formation after encountering an obstacle, T p It represents the total time of the entire obstacle avoidance process from the formation change to the restoration of the original formation after the obstacle avoidance is completed.

[0126] like Figure 5 As shown in the figure, the obstacle avoidance strategy process of the unmanned vehicle formation is as follows:

[0127] The unmanned vehicle formation moves in the expected formation. While moving, the laser radar scans the surrounding environment to determine whether there are obstacles around. If there are no obstacles, the formation will move in the original formation. When the laser radar detects that there are obstacles around, it calculates the relative distance between the unmanned vehicle and the obstacle and the maximum distance that can be passed in the obstacle area. If λ≥1, it means that the maximum passable distance in the obstacle area is larger than the current width of the unmanned vehicle formation. At this time, the unmanned vehicle formation does not need to change the formation and can pass through the obstacle area in the original formation. If ζ≤λ≤1, it means that the maximum passable distance in the current obstacle area is less than the width of the unmanned vehicle formation and the degree to which the formation needs to be compressed is less than the maximum degree that can be compressed. It means that the obstacle area can be passed by the formation contraction method. The unmanned vehicle formation needs to shrink the formation width to λ times the current formation width to pass the obstacle area. If λ≤ζ, it means that the maximum passable distance in the current obstacle area is less than the current formation width and the degree to which the formation needs to be compressed is greater than the degree to which the formation can be compressed. At this time, the formation switching obstacle avoidance method should be selected to pass through the obstacle area. Assuming that there are multiple formations with different widths in the formation knowledge base, first select the switching formations that meet the conditions, calculate the optimal formation through various evaluation indicators, and then switch the formation to avoid obstacles. Then the formation controller outputs specific posture error information to adjust the formation structure to avoid obstacles. After the leading unmanned vehicle finds that the last unmanned vehicle has passed the obstacle area, it re-issues the control command to change the formation, restores the current obstacle avoidance formation to the original expected formation state, and continues to perform the given task. At this point, the entire formation obstacle avoidance process ends.

[0128] Finally, it should be noted that: although the present invention and its advantages have been described in detail above, it should be understood that various changes, substitutions and transformations can be made without exceeding the spirit and scope of the present invention as defined by the appended claims. Moreover, the scope of the present invention is not limited to the specific embodiments of the processes, devices, means, methods and steps described in the specification. It will be easily understood by those of ordinary skill in the art from the disclosure of the present invention that existing and future processes, devices, means, methods or steps that perform substantially the same functions as the corresponding embodiments described herein or obtain substantially the same results as the corresponding embodiments described herein can be used according to the present invention. Therefore, the appended claims are intended to include such processes, devices, means, methods or steps within their scope.

Claims

1. An unmanned vehicle formation obstacle avoidance method based on sliding mode control and evaluation index, characterized by: The method mainly includes three parts: unmanned vehicle formation control algorithm design, unmanned vehicle formation overall obstacle avoidance strategy design and formation recovery; A. Design of unmanned vehicle formation control algorithm: Establish an unmanned vehicle leader-follower formation model and design a fixed-time sliding mode formation controller. The fixed-time sliding mode controller can converge the tracking error of the unmanned vehicle formation to zero within a fixed time and does not depend on the initial value of the system. B. Design of overall obstacle avoidance strategy for unmanned vehicle formation: adopting the overall obstacle avoidance mode of formation, designing the overall obstacle avoidance strategy of formation, establishing the formation knowledge base, and designing the overall obstacle avoidance strategy of unmanned vehicle formation taking into account environmental constraints; establishing the formation knowledge base, designing three formation obstacle avoidance modes of zero formation transformation, formation contraction transformation and formation switching transformation, setting the performance evaluation index of formation transformation, and establishing the decision-making mechanism of dynamic formation transformation, so that the unmanned vehicle formation can make the optimal formation transformation choice under different obstacle environments, and the integrity of the formation is effectively guaranteed; C. Formation recovery: After the leading unmanned vehicle detects that the last unmanned vehicle has passed the obstacle area, it issues a formation recovery command. The unmanned vehicle readjusts its own position and restores to the desired formation before the formation change, and continues to perform the task; The specific steps of designing the unmanned vehicle formation control algorithm are as follows: A. Establish a pilot-follower formation model: The pilot unmanned vehicle i moves in a two-dimensional plane according to a given linear velocity v and a given angular velocity ω, and the following unmanned vehicle j only needs to maintain the desired relative distance and relative angle By following the lead unmanned vehicle, the desired formation can be formed. Through this control structure, the entire unmanned vehicle formation control problem can be transformed into the trajectory tracking problem of the lead unmanned vehicle by the follower unmanned vehicle. The kinematic model of each unmanned vehicle is shown in Equation 1: Subscript j represents the jth following unmanned vehicle, L r is the distance between the UGV center of mass and the wheelbase, q = (x, y, θ) T Represents the pose vector of the unmanned vehicle in the (x, O, y) coordinate system, Indicates that the unmanned vehicle is at the reference trajectory point The pose vector, L ij represents the relative distance between the following unmanned vehicle and the leading unmanned vehicle, ψ ij represents the relative turning angle between the following unmanned vehicle and the leading unmanned vehicle, θ represents the direction of the unmanned vehicle's movement, that is, the angle between the velocity vector v and the positive direction of the θ axis, which is also the heading angle of the unmanned vehicle; Through the operation of spatial geometric relations, the trajectory tracking error model of the unmanned vehicle formation can be obtained as follows: in, represents the lateral trajectory tracking error of the unmanned vehicle formation, represents the longitudinal trajectory tracking error of the unmanned vehicle formation, represents the trajectory tracking error of the unmanned vehicle formation; B. Design a fixed-time sliding mode controller so that the trajectory tracking error between the following unmanned vehicle and the leading unmanned vehicle Converge to the smallest neighborhood of the origin; The designed sliding surface and fixed-time sliding mode control law are: After a series of mathematical operations, the final fixed-time sliding mode control law can be obtained as: in, α and β are fixed parameters of the fixed-time sliding mode control law; The designed fixed-time sliding mode controller can enable the unmanned vehicle formation to follow the trajectory of the leading unmanned vehicle within a fixed time, and the trajectory tracking error of each follower unmanned vehicle can converge to zero quickly within a fixed time and does not depend on the initial value of the system. C. Establishing a formation knowledge base: The purpose of establishing a formation knowledge base is to enable the unmanned vehicle formation to directly select the appropriate formation from the established formation knowledge base for obstacle avoidance according to the formation transformation mode during the obstacle avoidance process. First, each unmanned vehicle is assigned a unique number to mark different unmanned vehicles, such as Q1 and Q2. The number of the leading unmanned vehicle is set to 1, and the numbers of the other following unmanned vehicles are 2, 3, ..., n, where n is the number of vehicles in the unmanned vehicle formation. A parameter matrix is ​​established to represent the position relationship between the unmanned vehicles. The expression relationship is shown in the following formula: Among them, the matrix H represents the parameter value of a formation in the unmanned vehicle formation knowledge base. sj Indicates the jth unmanned vehicle Q in this formation j The state value of sj In, h 1j represents the j-th unmanned vehicle Q j Number information, h 2j represents j unmanned vehicles Q j The number of the pilot unmanned vehicle Q1 to be tracked, h 3j Represents the driverless car Q j The expected relative distance from the pilot unmanned vehicle Q1, h 4j Indicates Q j Compared with the expected angle of Q1, by establishing a formation knowledge base, the unmanned vehicle formation can quickly and effectively switch formations when switching formations to avoid obstacles, thereby passing through obstacle areas.

2. The unmanned vehicle formation obstacle avoidance method based on sliding mode control and evaluation index according to claim 1, characterized in that: When establishing the pilot-follower formation model, the concepts of time-varying relative distance and time-varying relative angle are introduced into the model to optimize the accuracy of the following unmanned vehicle when following the pilot unmanned vehicle; a. Time-varying relative distance: The expected relative distance refers to the minimum distance that must be maintained between two unmanned vehicles while ensuring driving safety. During the actual driving process of the unmanned vehicles, the expected relative distance between the two unmanned vehicles is time-varying; the expected relative distance is related to the relative speed of the leading vehicle and the following vehicle, the speed of the leading vehicle, and the speed of the following vehicle, as shown in Formula 5: Where S is the desired distance between the two vehicles, vrel is the relative speed of the two vehicles, and v f Indicates the speed difference between the leading vehicle and the following unmanned vehicle; The speed of the leading car is given, and the speed of the following car is obtained by the control system. Considering that one of the purposes of designing the controller is to make the speed of the following car infinitely close to the leading car, Formula 5 can be simplified, that is, vrel is set to 0, and Formula 6 can be obtained: L d =(0.851×vl+1.6) (6) Where L d is the expected relative distance between the two unmanned vehicles, vl is the speed of the pilot vehicle; b. Time-varying relative angle: The unmanned vehicles do not travel in a straight line on the actual road. The trajectory of the road may be an arbitrary curve. When the pilot vehicle changes direction, the relative angle between the two unmanned vehicles also changes. The expected relative angle between the two unmanned vehicles is defined as the initial expected relative angle when the two unmanned vehicles just start to travel. The difference between the two is expressed as formula 7: in, and are the initial expected relative angles of the two unmanned vehicles respectively; assuming the bending radius is r, through the leader-follower formation model, we can get Equation 8: Among them, ω l represents the angular velocity of the pilot car; according to the positional relationship between the two cars, we can get equation 9: Substituting equations 7 and 8 into equation 9, we can deduce the desired relative angle As shown in formula 10: It can be seen from formula 10 that the expected relative angle is related to the expected relative distance, the speed of the pilot vehicle and the angular velocity, and it is a time-varying variable.

3. The unmanned vehicle formation obstacle avoidance method based on sliding mode control and evaluation index according to claim 1, characterized in that: The specific steps for designing the overall obstacle avoidance strategy for the unmanned vehicle formation are as follows: The overall obstacle avoidance of the formation mainly includes two methods: formation contraction obstacle avoidance and formation switching obstacle avoidance; formation contraction obstacle avoidance means that after the unmanned vehicle formation encounters an obstacle, it calculates the size of the obstacle area that can be passed and its own formation width. If the length of the obstacle area that can be passed is less than the length of its own formation and the width of the formation can pass through the obstacle area without collision after the formation is compressed, the unmanned vehicle formation continues to maintain the original formation topology, that is, the formation shape remains unchanged, and the original expected formation is reasonably compressed while ensuring that there is no collision between unmanned vehicles, and then the obstacle area is passed. After passing the obstacle area, the formation is restored to its original state; formation switching obstacle avoidance means that when the unmanned vehicle formation cannot pass through the obstacle area by formation contraction obstacle avoidance transformation, the formation topology is changed according to the formation transformation strategy, that is, the original formation shape is changed to pass through the obstacle area, and then the formation is restored to the original expected formation; A. Establishing evaluation indicators for formation contraction and obstacle avoidance: a1. Establish the concept of deformation scale of the unmanned vehicle external formation to describe the degree to which the width of the current unmanned vehicle formation needs to be compressed; Among them, λ represents the deformation scale of the unmanned vehicle formation, d o represents the actual passable environment width of the current obstacle area, and l represents the width of the current unmanned vehicle formation; a2. Establish the concept of compression scale to describe the allowable degree of formation expansion and contraction; Among them, ζ represents the compression scale, δ represents the safe distance between unmanned vehicles, and σ represents the minimum distance between two adjacent unmanned vehicles in the current formation; B. Establishing evaluation indicators for formation switching and obstacle avoidance: b1. Establish a formation transformation rate index to describe the positional geometric change between the transformed formation and the expected formation before the transformation in the process of the unmanned vehicle formation avoiding obstacles, and use M d express: Among them, X r represents the formation matrix after the formation transformation, X p represents the expected formation of the unmanned vehicle platoon before the transformation, L Δ =X p -X r , represents the change in the geometric relationship of the formation matrix; b2. Establish a formation convergence time ratio indicator to describe how much formation change time accounts for the total time of the unmanned vehicle formation's overall obstacle avoidance process, using D u express: Among them, T r represents the formation change time of the unmanned vehicle formation after encountering an obstacle, T p It represents the total time of the entire obstacle avoidance process from the formation change to the restoration of the original formation after the obstacle avoidance is completed; C. Establish a dynamic transformation decision-making mechanism for the overall obstacle avoidance of the formation: c1. The formation does not change: When λ≥1, it indicates that the maximum passable distance in the obstacle area is larger than the current width of the unmanned vehicle formation. At this time, the unmanned vehicle formation does not need to change its formation and can pass through the obstacle area in the original formation, and the stability of the formation is guaranteed; c2. Formation contraction to avoid obstacles: When ζ≤λ≤1, it means that the maximum travel distance in the current obstacle area is less than the width of the unmanned vehicle formation and the degree to which the formation needs to be compressed is less than the maximum degree to which it can be compressed, indicating that the obstacle area can be passed by the formation contraction method. The unmanned vehicle formation needs to contract the formation width to λ times the current formation width to pass the obstacle area. Compared with the formation switching obstacle avoidance, the formation contraction obstacle avoidance method can ensure the stability and integrity of the formation; c3. Formation switching to avoid obstacles: When λ≤ζ, it means that the maximum travel distance in the current obstacle area is less than the current formation width and the degree to which the formation needs to be compressed is greater than the degree to which the formation can be compressed; if the formation is compressed to avoid obstacles, collisions will occur between unmanned vehicles. At this time, the formation switching obstacle avoidance method should be selected to pass through the obstacle area. Assuming that there are multiple formations with different widths in the formation knowledge base, first select the switching formations that meet the conditions, calculate the optimal formation through various evaluation indicators, and then perform formation switching to avoid obstacles.

Images