Vehicle path tracking method, device, vehicle, electronic device and medium

Abstract

The present disclosure provides a vehicle path tracking method, device, vehicle, electronic device and medium, which can be used in the field of unmanned driving. The method includes: obtaining planned path information and current motion state information of the vehicle; based on at least two different path tracking algorithms, calculating according to the geometric relationship between the current motion state information and the planned path information, respectively obtaining at least two control information for controlling the motion state of the vehicle at a future moment; within a dynamic limit window of the vehicle's motion state, performing non-uniform sampling around at least two control information to obtain sampled control information; wherein the sampling density of the window area close to the control information is greater than the sampling density of the window area far from the control information; determining the target control information of the vehicle at a future moment according to the degree of deviation between the trajectory prediction result corresponding to the sampled control information and the planned path information; and controlling the vehicle to move at a future moment according to the target control information.

Description

Technical Field

[0001] The present disclosure relates to the technical field of vehicle path tracking control, and in particular to a vehicle path tracking method, device, vehicle, electronic equipment and medium. Background Art

[0002] In automatic / unmanned vehicle technology, the vehicle needs to be controlled to travel along a predetermined route. Therefore, the vehicle is required to have good comprehensive performance such as accuracy, robustness, and real-time performance when performing path tracking control.

[0003] However, among the existing path tracking solutions, some have simple calculation methods that can meet the real-time requirements of calculation, but the path tracking accuracy and robustness are poor; some have complex calculation models that can meet the accuracy requirements of path tracking, but the required computing resources are expensive and it is difficult to meet the real-time requirements. Summary of the invention

[0004] In order to solve the above technical problems or at least partially solve the above technical problems, the embodiments of the present disclosure provide a method, device, vehicle, electronic device and medium for vehicle path tracking to achieve comprehensive performance such as better accuracy, robustness and real-time performance.

[0005] In the first aspect, an embodiment of the present disclosure provides a method for vehicle path tracking. The above method includes: obtaining the planned path information and current motion state information of the vehicle; based on at least two different path tracking algorithms, calculating according to the geometric relationship between the above current motion state information and the above planned path information, respectively obtaining at least two control information for controlling the motion state of the above vehicle at a future moment; within the dynamic limit window of the motion state of the above vehicle, performing non-uniform sampling around the above at least two control information to obtain sampled control information; wherein the sampling density of the window area close to the above control information is greater than the sampling density of the window area far from the above control information; determining the target control information of the above vehicle at the above future moment according to the deviation degree between the trajectory prediction result corresponding to the above sampled control information and the above planned path information; and controlling the above vehicle to move at the above future moment according to the above target control information.

[0006] According to an embodiment of the present disclosure, the above-mentioned current motion state information includes: the current speed and the current steering wheel angle; the above-mentioned dynamic restriction window is located in the two-dimensional plane formed by the speed and the steering wheel angle, the window center position of the above-mentioned dynamic restriction window is: the two-dimensional coordinate point formed by the above-mentioned current speed and the above-mentioned current steering wheel angle, and the window boundary of the above-mentioned dynamic restriction window is: the boundary speed and boundary steering wheel angle of the above-mentioned vehicle at the above-mentioned future moment under the acceleration limit information corresponding to the speed and the steering wheel angle.

[0007] According to an embodiment of the present disclosure, according to the degree of deviation between the trajectory prediction result corresponding to the above-mentioned sampling control information and the above-mentioned planned path information, the target control information of the above-mentioned vehicle at the above-mentioned future moment is determined, including: according to the above-mentioned sampling control information, simulating the above-mentioned vehicle from the current motion state for a preset time period to obtain a trajectory prediction result; eliminating invalid trajectory prediction results that will touch obstacles from the above-mentioned trajectory prediction results to obtain a valid trajectory prediction result; calculating according to the above-mentioned valid trajectory prediction result and the above-mentioned planned path information to obtain a cost function value, and the above-mentioned cost function value is used to indicate the degree of deviation of the above-mentioned valid trajectory prediction result compared to the above-mentioned planned path information; and determining the target sampling control information corresponding to the valid trajectory prediction result with the smallest cost function value as the target control information of the above-mentioned vehicle at the above-mentioned future moment.

[0008] According to an embodiment of the present disclosure, a cost function value is obtained by performing calculations based on the above-mentioned effective trajectory prediction result and the above-mentioned planned path information, including: obtaining the target prediction position point information in the above-mentioned effective trajectory prediction result and the target position point information in the above-mentioned planned path information, the time at which the above-mentioned target prediction position point information and the above-mentioned target position point information are located is the same time; calculating the Euclidean distance between the above-mentioned target prediction position point information and the above-mentioned target position point information, and the above-mentioned Euclidean distance is used as the above-mentioned cost function value.

[0009] According to an embodiment of the present disclosure, the path tracking algorithm includes: Pure-Pursuit algorithm and Stanley algorithm; the control information includes: steering wheel angle control information, which is used to control the steering wheel angle;

[0010] When calculated based on the Pure-Pursuit algorithm, the steering wheel angle control information satisfies the following expression:

[0011]

[0012] Among them, δ p represents the first steering wheel angle control information obtained based on the Pure-Pursuit algorithm; L represents the wheelbase of the vehicle; α represents the angle between the vehicle heading vector and the vehicle forward-looking vector, and the vehicle forward-looking vector represents the vector from the origin of the vehicle coordinate system to the forward-looking point of the vehicle; k v represents the first proportionality factor used to calculate the foresight distance; v f Indicates the current speed of the vehicle; set the vehicle head direction to the positive direction of the X axis, the X axis rotates 90° counterclockwise to the positive direction of the Y axis, and counterclockwise rotation is positive, δ max Indicates the maximum left turning angle of the vehicle's steering wheel; -δ max Indicates the maximum right turning angle of the vehicle's steering wheel;

[0013] When calculated based on the Stanley algorithm, the steering wheel angle control information satisfies the following expression:

[0014]

[0015] Among them, δ s represents the second steering wheel angle control information obtained based on the Stanley algorithm; e ψ represents the orientation deviation, which is the angle between the vehicle body orientation and the tangent direction at the nearest position point in the planned path information; k Δ Represents the second proportional coefficient acting on the lateral deviation; e Δ represents the lateral deviation, which is the Euclidean distance between the center of the front wheel axle of the vehicle and the nearest position point in the planned path information; k s Indicates the low-speed adjustment coefficient to avoid excessive steering wheel turning at low speed; k d Indicates the high-speed adjustment coefficient to avoid excessive steering wheel turning at high speed.

[0016] According to an embodiment of the present disclosure, the parameters of the path tracking algorithm are pre-optimized parameters; the method further includes: pre-optimizing the parameters of the path tracking algorithm. The optimization of the parameters of the path tracking algorithm includes: assigning preset parameters to the at least two path tracking algorithms, controlling the motion of the vehicle, and obtaining a path tracking result; optimizing and adjusting the preset parameters of the at least two path tracking algorithms according to the following deviation of the planned path information according to the path tracking result, until the target parameters after optimization and adjustment make the following deviation less than a set threshold.

[0017] In a second aspect, an embodiment of the present disclosure provides a device for vehicle path tracking. The device includes: an information acquisition module, a control information calculation module, a data sampling module, a target control information determination module and a control module. The information acquisition module is used to obtain the planned path information and current motion state information of the vehicle. The control information calculation module is used to calculate based on at least two different path tracking algorithms according to the geometric relationship between the current motion state information and the planned path information, and obtain at least two control information for controlling the motion state of the vehicle at a future moment. The data sampling module is used to perform non-uniform sampling around the at least two control information within the dynamic limit window of the motion state of the vehicle to obtain sampled control information; wherein the sampling density of the window area close to the control information is greater than the sampling density of the window area far from the control information. The target control information determination module is used to determine the target control information of the vehicle at the future moment according to the deviation degree between the trajectory prediction result corresponding to the sampled control information and the planned path information. The control module is used to control the vehicle to move at the future moment according to the target control information.

[0018] In a third aspect, an embodiment of the present disclosure provides a vehicle. The vehicle stores a set of instruction sets, which are executed by the vehicle to implement the vehicle path tracking method described above, or the device including the vehicle path tracking described above.

[0019] In a fourth aspect, an embodiment of the present disclosure provides an electronic device. The electronic device includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus; the memory is used to store a computer program; and the processor is used to implement the vehicle path tracking method described above when executing the program stored in the memory.

[0020] In a fifth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the vehicle path tracking method described above is implemented.

[0021] The above technical solution provided by the embodiments of the present disclosure has at least some or all of the following advantages:

[0022] Since at least two different path tracking algorithms perform path tracking calculations based on the geometric relationship between the current motion state information and the planned path information, the amount of calculation is relatively small; however, while utilizing the advantages of the two path tracking algorithms, it is considered that the two different path tracking algorithms each have their own problems: the accuracy of path tracking is not high enough, the robustness is low, and there is also a disadvantage that the vehicle dynamics model is ignored, resulting in uneven motion; the technical solution of the embodiment of the present disclosure uses at least two control information calculated by at least two different path tracking algorithms as prior information for sampling control, and performs non-uniform sampling within the dynamic limit window of the vehicle's motion state according to the prior information to obtain sampling control information, and uses the sampled sampling control information to perform trajectory prediction, and determines the target control information for vehicle motion control at a future moment according to the degree of deviation between the trajectory prediction result and the planned path information. This organic combination effectively overcomes the shortcomings of the two path tracking algorithms, improves the accuracy and robustness of path tracking, and reduces the number of samples through non-uniform sampling, which correspondingly reduces the calculation overhead, and has high tracking accuracy, robustness, motion smoothness, real-time and other comprehensive performance, which can meet the requirements of high-precision path tracking and real-time calculation of low-cost unmanned vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0025] Figure 1 A flow chart of a vehicle path tracking method according to an embodiment of the present disclosure is schematically shown;

[0026] Figure 2 A schematic diagram of a dynamic restriction window according to an embodiment of the present disclosure is schematically shown;

[0027] Figure 3 The schematic diagram schematically shows a scene corresponding to the Pure-Pursuit algorithm as the path tracking algorithm according to an embodiment of the present disclosure;

[0028] Figure 4 A schematic diagram of a scene corresponding to the Stanley algorithm is schematically shown according to an embodiment of the present disclosure;

[0029] Figure 5 The detailed implementation flow chart of step S140 according to an embodiment of the present disclosure is schematically shown;

[0030] Figure 6 The following schematically shows an implementation scenario diagram corresponding to step S140 according to an embodiment of the present disclosure;

[0031] Figure 7 A structural block diagram schematically shows a device for tracking a vehicle path according to an embodiment of the present disclosure; and

[0032] Figure 8 The structural block diagram of the electronic device provided by the embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0034] The first exemplary embodiment of the present disclosure provides a method for vehicle path tracking. The method can be executed by a vehicle or by an electronic device that can communicate with the vehicle and has computing capabilities.

[0035] In some embodiments, the above-mentioned vehicles are, for example, unmanned / autonomous driving vehicles, including unmanned inspection vehicles, unmanned express delivery vehicles, unmanned food delivery vehicles, unmanned mining vehicles, unmanned rescue vehicles, etc.

[0036] In some embodiments, the electronic device may be a computing device, such as a desktop computer, a laptop computer, a tablet computer, a smart phone, a vehicle-mounted device, etc. These devices can communicate with the vehicle or are located in the vehicle and have a path tracking control application installed; or the computing device may be a server that can communicate with the vehicle and provide path tracking services.

[0037] Figure 1 The flowchart of a vehicle path tracking method according to an embodiment of the present disclosure is schematically shown.

[0038] Reference Figure 1 As shown, the vehicle path tracking method provided by the embodiment of the present disclosure includes the following steps: S110, S120, S130, S140 and S150.

[0039] In step S110, the planned path information and current motion state information of the vehicle are obtained.

[0040] The above-mentioned vehicle may be a vehicle supporting automatic / unmanned driving functions. For example, in some embodiments, the planned path information of the vehicle may be provided by a vehicle path planner or other computing devices. In some embodiments, the current motion state information of the vehicle may be obtained through an on-board sensor, or the current motion state information of the vehicle may be obtained by monitoring the vehicle controller.

[0041] The planned path of the vehicle is a path connected by a number of sequentially arranged position points, including the position information of the sequentially arranged position points, and the planned path information can be used as navigation information of the vehicle. The planned path information is information presented based on a constructed map. In some embodiments, the content presented on the map is dynamically updated as the vehicle travels. The planned path information can be dynamically presented based on the vehicle coordinate system, or based on the ground coordinate system of the map itself.

[0042] In some embodiments, the position information of each position point in the planned path includes but is not limited to: the X coordinate of the position point in the reference coordinate system, the Y coordinate of the position point in the reference coordinate system, the orientation of the position point in the reference coordinate system (referring to the vector of the tangent at the position point pointing to the far end), the curvature of the planned path at the path point, etc. The reference coordinate system can be a ground coordinate system or a vehicle coordinate system or other coordinate systems that can be converted with these two coordinate systems.

[0043] The current motion state information of the vehicle is the instantaneous state information of the vehicle at the current moment, including but not limited to: the current speed v of the vehicle f and the current steering wheel angle δ f In some embodiments, for example, the reference coordinate system is a ground coordinate system, and the current motion state information further includes: position information (x, y, θ), where x and y correspond to the X coordinate and Y coordinate of the vehicle in the reference coordinate system, respectively, and θ represents the heading angle of the vehicle (representing the angle between the vehicle's center of mass speed and the horizontal axis X of the ground coordinate system).

[0044] The speed of a vehicle refers to the speed at which the vehicle is moving. The steering wheel angle of the vehicle is used to control the rotation of the vehicle tires, and then control the steering of the vehicle body. Different models have different corresponding front and rear wheel drive modes, and the range of the front and rear wheel angles corresponding to the steering wheel angle is also different.

[0045] In step S120, based on at least two different path tracking algorithms, calculations are performed according to the geometric relationship between the current motion state information and the planned path information to obtain at least two control information for controlling the motion state of the vehicle at a future moment.

[0046] The current motion state information and the planned path information are input into a path tracking algorithm (which can also be understood as a model) for calculation, and at least two types of control information for controlling the motion state of the above-mentioned vehicle at a future moment can be output.

[0047] The future moment may be a moment corresponding to the future 1s, the future 2s, the future 3s, etc. The shorter the time interval between the future moment and the current moment, the higher the control accuracy and the higher the granularity of the path tracking. The number of future moments may be one or more, and for each future moment, at least two corresponding control information needs to be obtained.

[0048] The at least two different path tracking algorithms mentioned above all perform path tracking calculations based on the geometric relationship between the current motion state information and the planned path information, and have the advantage of less computational complexity, that is, they can use a lightweight algorithm model to obtain at least two types of control information for controlling the motion state of the vehicle at future moments.

[0049] By performing path tracking calculations based on the geometric relationship between the current motion state information and the planned path information, the amount of calculation is small, but the accuracy is relatively low compared to the existing complex path tracking models.

[0050] In step S130, non-uniform sampling is performed around the at least two control information within a dynamic limit window of the motion state of the vehicle to obtain sampled control information; wherein the sampling density of the window area close to the control information is greater than the sampling density of the window area far from the control information.

[0051] While utilizing the advantages of at least two path tracking algorithms in step S120, the two different path tracking algorithms each have their own shortcomings: the accuracy of path tracking is not high enough, the robustness is low, and there is also a disadvantage of ignoring the vehicle dynamics model resulting in uneven motion; therefore, by executing steps S130 and S140, at least two control information calculated by at least two different path tracking algorithms are used as prior information for sampling control, and non-uniform sampling is performed within the dynamic limit window of the vehicle's motion state according to the prior information to obtain sampling control information, and trajectory prediction is performed using the sampled sampling control information, and the target control information used for vehicle motion control at future times is determined by the degree of deviation between the trajectory prediction result and the planned path information. This organic combination effectively overcomes the shortcomings of the two path tracking algorithms, improves the accuracy and robustness of path tracking, and reduces the number of samples through non-uniform sampling, which correspondingly reduces the computational overhead.

[0052] In step S130, the dynamic limit window of the vehicle's motion state refers to a state change range window of the vehicle's motion state corresponding to the current moment to the future moment, which is limited by the vehicle's acceleration information.

[0053] Figure 2 A schematic diagram of a dynamic restriction window according to an embodiment of the present disclosure is schematically shown.

[0054] According to the embodiments of the present disclosure, referring to Figure 2 As shown, the above current motion state information includes: current speed and current steering wheel angle. For example, in the motion state information of the vehicle at the current moment, Figure 2 In the figure, the current speed (speed in the driving direction) is represented as v1 (the unit is, for example, m / s), and the current steering wheel angle is represented as δ1 (the unit is, for example, angle or radian rad).

[0055] The dynamic limit window is located in a two-dimensional plane formed by the speed and the steering wheel angle (also described as the steering wheel angle). The center position of the dynamic limit window is: the two-dimensional coordinate point (v1, δ1) formed by the current speed and the current steering wheel angle. Figure 2 A hollow circle is used to indicate the center position of the window. It can be understood that the center position of the window can be displayed or not displayed in the two-dimensional plane.

[0056] The window boundary of the dynamic restriction window is: the boundary speed and boundary steering wheel angle of the vehicle at the future moment under the acceleration restriction information corresponding to the speed and the steering wheel angle.

[0057] For example, the maximum acceleration limit of a vehicle is a (in m / s 2 ), the maximum acceleration limit of the steering wheel is b (unit is, for example, angle / s 2 , or rad / s 2 ); then the window boundaries of the dynamic limit window of the motion state corresponding to the moment after 1 second in the future (i.e., an example of the future moment) are respectively: v1±a, δ1±b in two dimensions, wherein, when v1-a in v1±a, v1-a≥0 must be satisfied. Figure 2 In the figure, the horizontal axis indicates the steering wheel angle, the vertical axis indicates the speed, and the rectangular frame indicates the control space 200. In the right-hand coordinate system, the vehicle driving plane is the XOY plane, the direction perpendicular to the driving plane is the Z axis, the vehicle head direction is set as the positive direction of the X axis, the X axis rotates 90° counterclockwise to the positive direction of the Y axis, and the counterclockwise rotation is the positive direction of the angle. Figure 2 As shown, the right boundary of the control space 200 is the maximum turning angle δ of the vehicle steering wheel turning left (counterclockwise). max, the left boundary is the maximum turning angle of the vehicle steering wheel turning right (clockwise) -δ max The lower boundary of the control space 200 is 0, and the upper boundary is the limit value v of the vehicle speed. max The left and right boundaries of the dynamic limit window corresponding to the future moment 1 second after the current moment correspond to: δ1-b, δ1+b, and the upper and lower boundaries of the dynamic limit window correspond to: v1+a, v1-a, respectively.

[0058] Reference Figure 2 As shown, hollow circles are used to indicate control information, for example, two-dimensional coordinate points (v 21 , δ 21 ) and (v 22 , δ 22 ) to illustrate the two control information obtained based on two different path tracking algorithms. Figure 2 Solid dots are used to indicate sampling control information, see Figure 2 As shown, the sampling density of the window area close to the above control information is greater than the sampling density of the window area far from the above control information.

[0059] In some embodiments, for example, within the dynamic restriction window, the area is divided according to the distance from the at least two control information, for example, according to a preset distance, and divided into multiple areas from the inside to the outside, and each area uses a consistent sampling density, and different areas use different sampling densities according to the distance, and the closer the window area is to the control information, the greater the sampling density.

[0060] In step S140, the target control information of the vehicle at the future moment is determined according to the degree of deviation between the trajectory prediction result corresponding to the sampled control information and the planned path information.

[0061] In some embodiments, the sampling control information corresponding to the trajectory prediction result with the relatively lowest deviation degree may be used as the target control information at a future time (which may be one future time or multiple future time).

[0062] In step S150, the vehicle is controlled to move at the future time according to the target control information.

[0063] In the embodiment including the above steps S110 to S150, since at least two different path tracking algorithms perform path tracking calculations based on the geometric relationship between the current motion state information and the planned path information, the amount of calculation is relatively small; however, while utilizing the advantages of the two path tracking algorithms, it is considered that the two different path tracking algorithms each have the following disadvantages: the accuracy of path tracking is not high enough, the robustness is low, and there is also a disadvantage that the vehicle dynamics model is ignored, resulting in uneven motion; the technical solution of the embodiment of the present disclosure uses at least two control information calculated by at least two different path tracking algorithms as prior information for sampling control, and according to the prior information, in the motion of the vehicle Non-uniform sampling is performed within the dynamic limit window of the state to obtain sampling control information, and the sampling control information obtained by sampling is used to perform trajectory prediction. The target control information used for vehicle motion control in the future is determined by the degree of deviation between the trajectory prediction result and the planned path information. This organic combination effectively overcomes the shortcomings of the two path tracking algorithms, improves the path tracking accuracy and robustness, and reduces the number of samples through non-uniform sampling, which correspondingly reduces the computational overhead. It has high comprehensive performance such as tracking accuracy, robustness, motion smoothness and real-time performance, and can meet the requirements of high-precision path tracking and real-time calculation of low-cost unmanned vehicles.

[0064] According to an embodiment of the present disclosure, the above-mentioned path tracking algorithm includes: Pure-Pursuit algorithm and Stanley algorithm; the above-mentioned control information includes: steering wheel angle control information, which is used to control the size of the steering wheel angle.

[0065] Figure 3 A schematic diagram of a scene corresponding to a Pure-Pursuit algorithm is schematically shown according to an embodiment of the present disclosure.

[0066] Reference Figure 3 As shown, the foresight point and foresight distance in the planned path R of the vehicle are illustrated. The vehicle in this embodiment is an autonomous driving vehicle, and the Pure-Pursuit algorithm is a speed-based foresight distance adaptive algorithm. When the Pure-Pursuit algorithm is used for calculation, the steering wheel angle control information satisfies the following expression:

[0067]

[0068] Among them, δ p represents the first steering wheel angle control information obtained based on the Pure-Pursuit algorithm; L represents the wheelbase of the vehicle; α represents the vehicle heading vector (in Figure 3 The vehicle's forward sight vector (in Figure 3The vehicle forward sight vector represents the vector from the origin of the vehicle coordinate system to the forward sight point of the vehicle; k v represents the first proportionality factor used to calculate the foresight distance; v f Indicates the current speed of the vehicle. Figure 3 The linear speed is described in (here the speed along the vehicle's travel direction is described as the linear speed); the vehicle head direction is set to the positive direction of the X-axis, the X-axis rotates 90° counterclockwise to the positive direction of the Y-axis, and the counterclockwise rotation is positive, δ max Indicates the maximum left turning angle of the vehicle's steering wheel; -δ max Indicates the maximum right turning angle of the vehicle's steering wheel.

[0069] Among them, the coordinates (x, y) of the foresight point in the vehicle coordinate system can be obtained according to the vehicle positioning system and satisfy the following expression:

[0070]

[0071] Among them, d l Indicates the foresight distance.

[0072] exist Figure 3 Also illustrated is the instantaneous circular center, instantaneous circular motion radius and instantaneous trajectory of the vehicle moving along the first steering wheel angle control information.

[0073] One disadvantage of using the Pure-Pursuit algorithm alone for path tracking control is that it does not directly act on the orientation deviation. Another disadvantage is that when the current viewpoint is unavailable (when the end of the path is about to be reached), the controller will produce unstable movements.

[0074] Figure 4 A schematic diagram of a scene corresponding to the Stanley algorithm is schematically shown according to an embodiment of the present disclosure.

[0075] Reference Figure 4 As shown in the figure, the planned path of the vehicle is shown. The Stanley algorithm uses the lateral deviation (e Δ ) and orientation deviation (e ψ ) generates steering wheel angle control instructions.

[0076] When calculated based on the Stanley algorithm, the steering wheel angle control information satisfies the following expression:

[0077]

[0078] Among them, δ s represents the second steering wheel angle control information obtained based on the Stanley algorithm; e ψRepresents the orientation deviation, which is the tangent direction of the vehicle body orientation and the closest position point in the planned path information (in Figure 4 The angle between the tangent line of the nearest position point in the planned path is parallel to the dashed line in the figure; k Δ Represents the second proportional coefficient acting on the lateral deviation; e Δ represents the lateral deviation, which is the Euclidean distance between the center of the vehicle's front wheel axle and the nearest position point in the planned path information; v f Indicates the current speed of the vehicle. Figure 4 is described as linear speed (here the speed along the vehicle's travel direction is described as linear speed); k s Indicates the low-speed adjustment coefficient to avoid excessive steering wheel turning at low speed (for example, when the car speed is close to 0); k d Indicates the high-speed adjustment coefficient to avoid excessive steering wheel turning at high speed.

[0079] Among them, k s It is a coefficient to avoid numerical instability and excessive steering wheel angle when the car speed is very low (speed close to 0); k d It is used to reduce the steering wheel angle when the car is at a high speed to avoid over-turning.

[0080] The Stanley algorithm is used alone for path tracking control. Since it uses the position point closest to the vehicle in the planned path to calculate the lateral error, and it acts directly on the heading error, these two facts will make it react aggressively at low speeds (if the parameters are adjusted for high speeds) or react sluggishly at high speeds (if the parameters are adjusted for low speeds).

[0081] According to an embodiment of the present disclosure, the parameters of the path tracking algorithm are pre-optimized parameters. The method further includes: pre-optimizing the parameters of the path tracking algorithm. For example, the parameters of the Pure-Pursuit algorithm and the Stanley algorithm are pre-optimized so that the target parameters after optimization and adjustment make the path tracking results obtained by the two algorithms follow the planned path information of the vehicle less than a set threshold. The set threshold can be adjusted according to the accuracy requirements of the path tracking.

[0082] In one embodiment, the parameters of the above-mentioned path tracking algorithm are optimized, including: assigning preset parameters to the above-mentioned at least two path tracking algorithms, performing motion control on the above-mentioned vehicle, and obtaining a path tracking result; optimizing and adjusting the preset parameters of the above-mentioned at least two path tracking algorithms based on the following deviation of the above-mentioned planned path information according to the above-mentioned path tracking result, until the optimized and adjusted target parameters make the above-mentioned following deviation less than a set threshold.

[0083] For example, in some embodiments, the first proportionality coefficient k v It is an adjustable parameter of the Pure-Pursuit algorithm. By observing the path tracking effect of the vehicle under given parameters, the parameter is increased or decreased until the vehicle can follow the path with high accuracy. The tuning basis is: if the value is too large, the path tracking will be inaccurate; if the value is too small, the path tracking is prone to oscillation.

[0084] In some embodiments, the second proportionality coefficient k Δ , low speed adjustment coefficient k s and high speed adjustment coefficient k d is an adjustable parameter of the Stanley algorithm. The second proportionality coefficient k Δ The parameter is obtained by debugging on the vehicle. By observing the vehicle path tracking effect under given parameters, the parameter is increased or decreased until the vehicle can follow the path with high accuracy. The debugging method can be: if the vehicle steering wheel angle response is relatively slow, increase the parameter; if the vehicle steering wheel angle is excessive, decrease the parameter. Low speed adjustment coefficient k s The actual debugging method includes: let the vehicle run at a speed close to 0, and then s = 0, the steering wheel will turn too far, and you need to increase k slowly. s This parameter can be obtained when the steering wheel angle is appropriate and no longer excessive. High-speed adjustment coefficient k d Debugging principles: To avoid excessive steering wheel angle when the vehicle is driving at high speed, a damping coefficient needs to be added. The higher the driving speed, the smaller the steering wheel angle needs to be. The actual debugging methods include: Let the vehicle drive at a very high speed, and adjust the k d This parameter is obtained by starting from 1 and slowly increasing it until the steering wheel angle is conservative when driving at high speed.

[0085] By optimizing the parameters of the path tracking algorithm in advance, it is ensured that the at least two control information corresponding to the at least two different path tracking algorithms used in step S120 are relatively reliable, that is, the reliability and certain accuracy of the prior information are guaranteed.

[0086] In some embodiments, the control information includes not only steering wheel angle control information but also speed control information for controlling the speed.

[0087] In some embodiments, the speed control information is the target linear speed that the vehicle wants to achieve. The target linear speed is usually determined based on the specified linear speed information (for example, the specified speed for urban roads is 40 km / h, and the specified speed for highways is 110 km / h), the driving strategy information of the path complexity (for example, acceleration on straight roads and deceleration on curves), the road condition information (for example, driving at a low speed if there are pedestrians around, and driving at a high speed if the surroundings are empty), and the linear acceleration constraints.

[0088] The vehicle speed can then be tracked by a longitudinal controller, such as a commonly used PID (proportional-integral-derivative) controller, which generates a throttle control command for the vehicle based on the desired target linear speed and the actual linear speed.

[0089] Figure 5 The detailed implementation flow chart of step S140 according to an embodiment of the present disclosure is schematically shown; Figure 6 The following schematically shows an implementation scenario diagram corresponding to step S140 according to an embodiment of the present disclosure.

[0090] According to the embodiments of the present disclosure, referring to Figure 5 As shown, in the above step S140, the target control information of the above vehicle at the above future moment is determined according to the degree of deviation between the trajectory prediction result corresponding to the above sampling control information and the above planned path information, including the following steps: S510, S520, S530 and S540.

[0091] In step S510, based on the sampling control information, the vehicle is simulated from the current motion state for a preset time period to obtain a trajectory prediction result.

[0092] The duration of the preset time period may be greater than or equal to the duration from the current moment to the future moment.

[0093] By performing forward simulation on each sampled control information, a trajectory prediction result is obtained. During the preset time period, the control command remains constant, that is, the vehicle is always controlled to move with the sampled control information (for example, with constant steering wheel angle information and linear speed information) during the preset time period. Figure 6 As shown in Figure 1, due to the large number of sampling points of the sampling control information, a set of trajectory prediction results will be generated accordingly. Figure 6 The dotted line extending from the origin of the vehicle coordinate system is used to illustrate the trajectory prediction result.

[0094] In step S520, invalid trajectory prediction results that may hit obstacles are eliminated from the above trajectory prediction results to obtain valid trajectory prediction results.

[0095] By eliminating invalid trajectory prediction results that may touch obstacles, the remaining trajectory prediction results are valid trajectory prediction results.

[0096] In step S530, a cost function value is obtained by calculating the effective trajectory prediction result and the planned path information, where the cost function value is used to indicate the degree of deviation of the effective trajectory prediction result from the planned path information.

[0097] In some embodiments, in the above step S530, a cost function value is obtained by performing calculations based on the above effective trajectory prediction results and the above planned path information, including: obtaining the target prediction position point information in the above effective trajectory prediction results and the target position point information in the above planned path information, the time at which the above target prediction position point information and the above target position point information are located is the same time; calculating the Euclidean distance between the above target prediction position point information and the above target position point information, and the above Euclidean distance is used as the above cost function value.

[0098] For example, refer to Figure 6 As shown in the figure, the planned path R, the foresight point and the foresight distance are illustrated. The foresight point in the planned path is taken as the target position point, and the target prediction position point information at the same time as the foresight point can be obtained from multiple valid trajectory prediction results; the cost function value is obtained by calculating the Euclidean distance between the target prediction position point and the target position point.

[0099] By taking the foresight point as the target position point, the cost function of the optimization problem can be simplified, the computing power overhead of solving the optimization problem can be reduced, and high path tracking accuracy can be achieved.

[0100] In step S540, the target sampling control information corresponding to the effective trajectory prediction result with the minimum cost function value is determined as the target control information of the vehicle at the future moment.

[0101] In step S140, the control strategy of model predictive control is adopted, which treats the control task as a constrained optimization problem. The future state of the vehicle is predicted within the prediction range by using the vehicle motion model, and then a constrained online optimization problem is solved to select the optimal control input that minimizes the cost function. The cost function value can be the degree of alignment between the future state of the vehicle and the target state, whether the future state of the vehicle has the risk of colliding with obstacles, etc. For example, given the current state of the vehicle and the target path for tracking, the model predictive control uses the vehicle motion model to simulate different control inputs within a certain prediction range (these inputs are not actually applied to the vehicle) to predict the resulting future state (obtain a trajectory prediction result, which can be a predicted trajectory). At each time step, the optimal control input set corresponding to the trajectory with the smallest cost function value is selected (considering the constraints), and the front part of the optimal control input set (within a certain control range) is applied to the vehicle, and the remaining part is discarded. As the vehicle state is updated, a new optimal predicted trajectory is calculated within a certain prediction range by repeating the same algorithm, and a new optimal control input is calculated.

[0102] In the above embodiment, by organically combining the control strategy of model predictive control with a priori input, the a priori input is the control information of the vehicle calculated by using the geometric relationship between the vehicle kinematic model and the planned path, effectively overcoming the shortcomings of the two path tracking algorithms, the Pure-Pursuit algorithm and the Stanley algorithm, retaining the advantage of low computational complexity, and improving the accuracy and robustness of path tracking, and also reducing the number of samples through non-uniform sampling, correspondingly reducing the computing power overhead of forward simulation, and sampling within the dynamic window limited by the vehicle acceleration (linear acceleration and steering wheel angular acceleration) constraint, that is, reflecting the vehicle dynamics constraint, and applying the vehicle dynamics constraint when sampling within the dynamic window limited by the acceleration constraint in the control space, compared with the simple Pure-Pursuit controller or Stanley controller, the motion control effect is smoother. It has both high tracking accuracy, robustness, motion smoothness and real-time comprehensive performance, and can meet the requirements of high-precision path tracking and real-time calculation of low-cost unmanned vehicles.

[0103] A second exemplary embodiment of the present disclosure provides a vehicle path tracking device, which may be a vehicle-mounted device installed on a vehicle, or a device independent of the vehicle and capable of communicating with the vehicle.

[0104] Figure 7 The structure block diagram of the vehicle path tracking device according to the embodiment of the present disclosure is schematically shown.

[0105] Reference Figure 7As shown, the vehicle path tracking device 700 provided by the embodiment of the present disclosure includes: an information acquisition module 701, a control information calculation module 702, a data sampling module 703, a target control information determination module 704 and a control module 705.

[0106] The above-mentioned information acquisition module 701 is used to obtain the planned path information and current motion state information of the vehicle.

[0107] The control information calculation module 702 is used to perform calculations based on at least two different path tracking algorithms according to the geometric relationship between the current motion state information and the planned path information, and obtain at least two control information for controlling the motion state of the vehicle at a future moment.

[0108] The data sampling module 703 is used to perform non-uniform sampling around the at least two control information within the dynamic limit window of the motion state of the vehicle to obtain sampled control information; wherein the sampling density of the window area close to the control information is greater than the sampling density of the window area far from the control information.

[0109] The target control information determination module 704 is used to determine the target control information of the vehicle at the future moment according to the degree of deviation between the trajectory prediction result corresponding to the sampled control information and the planned path information.

[0110] The control module 705 is used to control the vehicle to move at the future time according to the target control information.

[0111] According to an embodiment of the present disclosure, the above-mentioned path tracking algorithm includes: Pure-Pursuit algorithm and Stanley algorithm; the above-mentioned control information includes: steering wheel angle control information, which is used to control the size of the steering wheel angle.

[0112] According to an embodiment of the present disclosure, the parameters of the path tracking algorithm are pre-optimized parameters. The apparatus 700 further includes: a parameter optimization module, which is used to pre-optimize the parameters of the path tracking algorithm.

[0113] For example, the parameter optimization module is used to optimize the parameters of the Pure-Pursuit algorithm and the Stanley algorithm in advance, so that the target parameters after optimization and adjustment make the path tracking results obtained by the two algorithms follow the planned path information of the vehicle less than a set threshold. The set threshold can be adjusted according to the accuracy requirements of the path tracking.

[0114] In some embodiments, the control information includes not only steering wheel angle control information but also speed control information for controlling the speed.

[0115] The target control information determination module 704 includes: a forward simulation submodule, an invalid result elimination submodule, a cost function calculation submodule, and a target control information determination submodule.

[0116] The forward simulation submodule is used to simulate the vehicle from the current motion state for a preset time period according to the sampling control information to obtain a trajectory prediction result.

[0117] The invalid result elimination submodule is used to eliminate the invalid trajectory prediction results that may touch obstacles from the trajectory prediction results to obtain valid trajectory prediction results.

[0118] The cost function calculation submodule is used to calculate according to the effective trajectory prediction result and the planned path information to obtain a cost function value, and the cost function value is used to indicate the degree of deviation of the effective trajectory prediction result compared with the planned path information.

[0119] The target control information determination submodule is used to determine the target sampling control information corresponding to the effective trajectory prediction result with the minimum cost function value as the target control information of the vehicle at the future moment.

[0120] Any multiple of the functional modules or submodules included in the above-mentioned device 700 can be combined in one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. At least one of the functional modules or submodules included in the above-mentioned device 700 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in any appropriate combination of any of them. Alternatively, at least one of the functional modules or submodules included in the above-mentioned device 700 can be at least partially implemented as a computer program module, and when the computer program module is run, the corresponding function can be executed.

[0121] A third exemplary embodiment of the present disclosure provides a vehicle. The vehicle stores a set of instruction sets, which are executed by the vehicle to implement the vehicle path tracking method described above, or the device including the vehicle path tracking described above.

[0122] The vehicle may be, for example, a vehicle that supports automatic / unmanned driving functions, and can realize automatic control of vehicle driving, including control of the vehicle steering wheel angle and throttle to realize control of vehicle speed and steering, etc. Including: unmanned inspection vehicles, unmanned express delivery vehicles, unmanned food delivery vehicles, unmanned mining vehicles, unmanned rescue vehicles, etc.

[0123] A fourth exemplary embodiment of the present disclosure provides an electronic device.

[0124] Figure 8 The structural block diagram of the electronic device provided by the embodiment of the present disclosure is schematically shown.

[0125] Reference Figure 8 As shown, the electronic device 800 provided by the embodiment of the present disclosure includes a processor 801, a communication interface 802, a memory 803 and a communication bus 804, wherein the processor 801, the communication interface 802 and the memory 803 communicate with each other through the communication bus 804; the memory 803 is used to store computer programs; the processor 801 is used to implement the vehicle path tracking method as described above when executing the program stored in the memory.

[0126] The fifth exemplary embodiment of the present disclosure further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the vehicle path tracking method described above is implemented.

[0127] The computer-readable storage medium may be included in the device / apparatus described in the above embodiment; or it may exist independently without being assembled into the device / apparatus. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0128] According to an embodiment of the present disclosure, a computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0129] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0130] The foregoing is merely a specific embodiment of the present disclosure, which enables those skilled in the art to understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features applied herein.

Claims

1. A vehicle path tracking method, characterized in that: include: Obtain the vehicle's planned path information and current motion status information; Based on at least two different path tracking algorithms, calculation is performed according to the geometric relationship between the current motion state information and the planned path information to obtain at least two control information for controlling the motion state of the vehicle at a future moment; The path tracking algorithms include: Pure-Pursuit algorithm and Stanley algorithm; In the dynamic limit window of the motion state of the vehicle, non-uniform sampling is performed around the at least two control information to obtain sampled control information; wherein the sampling density of the window area close to the control information is greater than the sampling density of the window area far from the control information; the at least two control information are used as prior information for sampling control; the dynamic limit window of the motion state of the vehicle refers to a state change range window of the motion state corresponding to the vehicle from the current moment to the future moment, which is limited by the acceleration information corresponding to the vehicle speed and the steering wheel angle; Determining target control information of the vehicle at the future time according to a degree of deviation between a trajectory prediction result corresponding to the sampled control information and the planned path information; According to the target control information, the vehicle is controlled to move at the future time.

2. The method according to claim 1, characterized in that The current motion state information includes: current speed and current steering wheel angle; The dynamic limitation window is located in a two-dimensional plane formed by the speed and the steering wheel angle. The window center position of the dynamic limitation window is: a two-dimensional coordinate point formed by the current speed and the current steering wheel angle. The window boundary of the dynamic limitation window is: under the acceleration limitation information corresponding to the speed and the steering wheel angle, the boundary speed and boundary steering wheel angle of the vehicle at the future moment.

3. The method according to claim 1, characterized in that Determining target control information of the vehicle at the future time according to the deviation degree between the trajectory prediction result corresponding to the sampled control information and the planned path information, including: According to the sampling control information, the vehicle is simulated for a preset time period from a current motion state to obtain a trajectory prediction result; Eliminating invalid trajectory prediction results that may touch obstacles from the trajectory prediction results to obtain valid trajectory prediction results; Calculating according to the effective trajectory prediction result and the planned path information to obtain a cost function value, wherein the cost function value is used to indicate the degree of deviation of the effective trajectory prediction result compared to the planned path information; The target sampling control information corresponding to the effective trajectory prediction result with the minimum cost function value is determined as the target control information of the vehicle at the future moment.

4. The method according to claim 3, characterized in that The cost function value is obtained by calculating the effective trajectory prediction result and the planned path information, including: Acquire target predicted position point information in the effective trajectory prediction result and target position point information in the planned path information, wherein the target predicted position point information and the target position point information are at the same time; The Euclidean distance between the target predicted position point information and the target position point information is calculated, and the Euclidean distance is used as the cost function value.

5. The method according to claim 1, characterized in that The control information includes: steering wheel angle control information, which is used to control the steering wheel angle; When the calculation is performed based on the Pure-Pursuit algorithm, the steering wheel angle control information satisfies the following expression: Among them, δ p represents the first steering wheel angle control information obtained based on the Pure-Pursuit algorithm; L represents the wheelbase of the vehicle; α represents the angle between the vehicle heading vector and the vehicle forward-looking vector, and the vehicle forward-looking vector represents the vector from the origin of the vehicle coordinate system to the forward-looking point of the vehicle; k v represents the first proportionality factor used to calculate the foresight distance; v f Indicates the current speed of the vehicle; set the vehicle head direction to the positive direction of the X axis, the X axis rotates 90° counterclockwise to the positive direction of the Y axis, and counterclockwise rotation is positive, δ max Indicates the maximum left turning angle of the vehicle's steering wheel; -δ max Indicates the maximum right turning angle of the vehicle's steering wheel; When the calculation is performed based on the Stanley algorithm, the steering wheel angle control information satisfies the following expression: Among them, δ s represents the second steering wheel angle control information obtained based on the Stanley algorithm; e ψ represents the orientation deviation, which is the angle between the vehicle body orientation and the tangent direction at the nearest position point in the planned path information; k Δ Represents the second proportional coefficient acting on the lateral deviation; e Δ represents the lateral deviation, which is the Euclidean distance between the center of the front wheel axle of the vehicle and the nearest position point in the planned path information; k s Indicates the low-speed adjustment coefficient to avoid excessive steering wheel turning at low speed; k d Indicates the high-speed adjustment coefficient to avoid excessive steering wheel turning at high speed.

6. The method according to claim 1 or 5, characterized in that: The parameters of the path tracking algorithm are pre-optimized parameters; The method further includes: optimizing the parameters of the path tracking algorithm in advance; the optimizing the parameters of the path tracking algorithm includes: Assigning preset parameters to the at least two different path tracking algorithms, performing motion control on the vehicle, and obtaining a path tracking result; According to the following deviation of the planned path information from the path tracking result, the preset parameters of the at least two different path tracking algorithms are optimized and adjusted until the target parameters after optimization and adjustment make the following deviation less than a set threshold.

7. A vehicle path tracking device, characterized in that: include: An information acquisition module is used to obtain the planned path information and current motion state information of the vehicle; a control information calculation module, configured to calculate based on at least two different path tracking algorithms according to a geometric relationship between the current motion state information and the planned path information, and obtain at least two types of control information for controlling the motion state of the vehicle at a future moment; The path tracking algorithms include: Pure-Pursuit algorithm and Stanley algorithm; A data sampling module, for performing non-uniform sampling around the at least two control information within a dynamic limit window of the motion state of the vehicle to obtain sampled control information; wherein the sampling density of the window area close to the control information is greater than the sampling density of the window area far from the control information; the at least two control information are used as prior information for sampling control; the dynamic limit window of the motion state of the vehicle refers to a state change range window of the motion state corresponding to the vehicle from the current moment to the future moment, which is limited by the acceleration information corresponding to the vehicle speed and the steering wheel angle; A target control information determination module, used to determine the target control information of the vehicle at the future moment according to the degree of deviation between the trajectory prediction result corresponding to the sampled control information and the planned path information; A control module is used to control the vehicle to move at the future moment according to the target control information.

8. A vehicle storing a set of instruction sets, wherein the instruction sets are executed by the vehicle to implement the method according to any one of claims 1 to 6, or the device according to claim 7.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; A processor, for implementing the method according to any one of claims 1 to 6 when executing a program stored in a memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images