Unmanned vehicle path tracking method

Abstract

The invention discloses an unmanned vehicle path tracking method, which comprises the following steps of: 1, establishing a vehicle simplified kinematics model in a vehicle driving process; step 2, based on a genetic algorithm, calculating a membership function of fuzzy control by taking the speed and the lateral deviation as input and taking the foresight distance of the vehicle as output; and step 3, establishing a fuzzy controller, obtaining the foresight distance of the vehicle according to the deviation between the actual trajectory of the vehicle and the target trajectory and the vehicle running speed, and calculating the expected front wheel turning angle of the vehicle based on a tracking control method to complete the path tracking control of the vehicle. Path tracking control of the vehicle is achieved by calculating the expected front wheel turning angle of the vehicle.

Description

TECHNICAL FIELD

[0001] The application relates to an unmanned vehicle path tracking method and belongs to the technical research field of unmanned vehicles. BACKGROUND

[0002] With the rapid progress of intelligent driving technology, unmanned vehicles have become the focus of research in various fields and are gradually penetrating various commercial fields. The main key technologies include environment perception, path planning, decision-making and control technology.

[0003] For path tracking control research, methods can be divided into two categories according to whether a vehicle model is established. The model-based control method includes pure tracking control method, Stanley control algorithm, LQR and MPC. This method is relatively simple to implement, but the calculation accuracy will decrease as the complexity of the model increases. The model-free control method includes PID control and neural network method. This method requires a large number of training samples and a long training time. A path tracking control method based on the geometric model of the vehicle is called pure tracking control. According to the deviation of the actual position and the ideal position of the vehicle, as well as the deviation of the actual turning angle and the ideal turning angle, the method adjusts the feedback to correct the errors, thereby achieving good control effect. This method is a proportional controller that can convert the longitudinal deviation of the position and the required position into lateral control. This strategy shows excellent robustness, and even in the case of large lateral error or discontinuous reference path curve, excellent tracking effect can be achieved. SUMMARY

[0004] The application designs and develops an unmanned vehicle path tracking method, which realizes path tracking control of the vehicle by calculating the expected front wheel turning angle of the vehicle.

[0005] The technical scheme provided by the application is as follows:

[0006] An unmanned vehicle path tracking method comprises the following steps:

[0007] Step 1: During the driving process of the vehicle, a simplified kinematic model of the vehicle is established.

[0008] Step 2: Based on the genetic algorithm, the speed and lateral deviation are taken as inputs, and the front-view distance of the vehicle is taken as output to calculate the membership function of the fuzzy control.

[0009] Step 3: A fuzzy controller is established. The front-view distance of the vehicle is obtained through the deviation between the actual trajectory and the target trajectory of the vehicle and the driving speed of the vehicle. The expected front wheel turning angle of the vehicle is calculated based on the tracking control method, and the path tracking control of the vehicle is completed.

[0010] The expected front wheel turning angle calculation formula of the vehicle is as follows:

[0011]

[0012] wherein, a(t) represents the included angle between the vehicle body and the target path point in unit sampling time.

[0013] Preferably, the vehicle simplified kinematics model is:

[0014]

[0015] wherein, v r is the speed of the vehicle, is the heading angle, l is the wheelbase of the front and rear wheels, δ f is the steering angle of the front wheel.

[0016] Preferably, the maximum value of the objective function optimized by the genetic algorithm is:

[0017]

[0018] wherein, J(x) is the optimal function for evaluating the objective function controller;

[0019] The calculation formula of the optimal function J(x) is:

[0020]

[0021] wherein, k is the duration, w1, w2, w3 are the weighting coefficients of e(t), ec(t), u(t) respectively.

[0022] Preferably, the membership function of the input quantity speed obtained based on the genetic algorithm in step two is selected from the combination of trapezoidal membership function and triangular membership function, the membership function of the input quantity lateral deviation is selected from the combination of Gaussian membership function and triangular membership function, and the membership function of the output quantity forward distance is selected from the combination of trapezoidal membership function and triangular membership function.

[0023] Preferably, the domain of the vehicle driving speed V is [0, 2m / s], the domain of the vehicle lateral deviation EX is [-0.6m, 0.6m], and the domain of the vehicle forward distance L d is [1m, 10m].

[0024] Preferably, the vehicle driving speed V is divided into 5 levels {very slow, slower, moderate, faster, very fast}, and the corresponding fuzzy set is {LD, LX, M, RX, RD}; the lateral deviation EX of the actual trajectory and the target trajectory of the vehicle is divided into 5 levels {right large, right small, center, left small, left large}, and the corresponding fuzzy set is {LD, LX, M, RX, RD}; and the forward distance L dThe five levels are {closer, near, middle, far, farther}, and the fuzzy set is {VN, N, M, F, VF}.

[0025] Preferably, the front view distance in step three includes a front view distance part of the fuzzy control output and a minimum front view distance part,

[0026] The minimum front view distance calculated based on the front view distance of the fuzzy control output is:

[0027] L0 = L d -k v v;

[0028] In the formula, L0 is the minimum front view distance, k v is the speed ratio coefficient, and v is the vehicle speed.

[0029] The beneficial effects of the present application: based on the genetic algorithm optimization membership function of fuzzy control strategy, combined with the pure tracking model, the optimal front wheel steering angle is calculated to meet the needs of unmanned vehicle path tracking, improve the efficiency and accuracy of unmanned vehicle path tracking.

[0030] The method can be applied to an unmanned cleaning vehicle in an airport, and systematically improves the detection accuracy and positioning reliability of the unmanned cleaning vehicle in the airport in a complex runway environment, while ensuring the stability of the model under variable weather conditions, and provides a high reliability identification basis for automatic removal of foreign objects in the airport. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 It is a schematic diagram of the vehicle kinematic model described in the present application.

[0032] Figure 2 It is a flowchart of the genetic optimization algorithm described in the present application.

[0033] Figure 3(a) is a schematic diagram of the vehicle speed membership function described in the present application.

[0034] Figure 3(b) is a schematic diagram of the vehicle lateral deviation membership function described in the present application.

[0035] Figure 4 It is a membership function diagram of the fuzzy control output variable described in the present application.

[0036] Figure 5 It is a schematic diagram of the pure tracking model described in the present application.

[0037] Figure 6 It is a lateral control tracking curve diagram described in the present application

[0038] Figure 7 It is a schematic diagram of the optimized front view distance described in the present application.

[0039] Figure 8 A lateral deviation variation curve comparison chart according to the present application.

[0040] Figure 9 A front wheel steering angle variation curve comparison chart according to the present application. DETAILED DESCRIPTION

[0041] The present application will be further described in detail below with reference to the accompanying drawings so that those skilled in the art can implement the present application according to the description and drawings.

[0042] As shown in Figures 1-9 , the present application provides an unmanned vehicle path planning and tracking method, which calculates an optimal front wheel steering angle to meet the needs of unmanned vehicle path tracking by a fuzzy control strategy based on genetic algorithm optimization of membership function, in combination with a pure tracking model, to improve the efficiency and accuracy of unmanned vehicle path tracking, comprising:

[0043] Step one, during vehicle driving, a simplified kinematic model of the vehicle is established according to the following formula:

[0044] The simplified kinematic model of the vehicle is as follows:

[0045]

[0046] In the formula, v r is the speed of the vehicle, is the heading angle, l is the wheelbase of the front and rear wheels, and δ f is the steering angle of the front wheel.

[0047] Step two, based on genetic algorithm, the speed and lateral deviation are taken as inputs, and the front view distance of the vehicle is taken as output to calculate the membership function of fuzzy control.

[0048] The genetic algorithm automatically designs the membership function parameters and rule table of the fuzzy controller by optimizing the objective function J(x), thereby establishing a dynamic mapping relationship from the driving speed and lateral deviation to the front view distance, and realizing adaptive optimization of the parameters of pure tracking control.

[0049] As shown in Figure 2 , the genetic algorithm process includes population initialization, population fitness calculation, genetic operation, and finally the membership function of fuzzy control.

[0050] The maximum value of the objective function optimized by the genetic algorithm is as follows:

[0051]

[0052] In the formula, J(x) is the optimal function for evaluating the objective function controller.

[0053] The formula of the optimal function J(x) is:

[0054]

[0055] In the formula, k is the time duration, w1, w2, w3 are the weighting coefficients of e(t), ec(t), u(t) respectively.

[0056] In the present application, as a preferred example, w1=0.5, w2=0.4, w3=0.1.

[0057] Table 1 Genetic algorithm related parameters

[0058]

[0059] The membership functions of the input variables of the fuzzy control obtained by the genetic optimization algorithm are shown in Fig. 3. The membership function of the input variable: speed is selected as a combination of trapezoidal membership function and triangular membership function, the membership function of the input variable: lateral deviation is selected as a combination of Gaussian membership function and triangular membership function, and the membership function of the output variable of the fuzzy control is shown in Fig. 4, in which the membership function of the output variable: look-ahead distance L d is selected as a combination of trapezoidal membership function and triangular membership function. Figure 4

[0060] The domain of the vehicle driving speed V in the fuzzy control obtained by the genetic algorithm is [0, 2m / s], the domain of the lateral deviation EX of the vehicle is [-0.6m, 0.6m], and the domain of the look-ahead distance L d of the vehicle is [1m, 10m].

[0061] The vehicle driving speed V is divided into 5 levels, and the fuzzy set is {LD, LX, M, RX, RD}; the lateral deviation EX of the actual trajectory and the target trajectory of the vehicle is divided into 5 levels, and the fuzzy set is {LD, LX, M, RX, RD}; and the look-ahead distance L d of the vehicle is divided into 5 levels, and the fuzzy set is {VN, N, M, F, VF}.

[0062] Step 3: Establishing the fuzzy controller, obtaining the look-ahead distance L d of the vehicle through the lateral deviation of the actual trajectory and the target trajectory of the vehicle and the driving speed of the vehicle, and calculating the expected front wheel angle of the vehicle in a pure tracking control method through the look-ahead distance L d of the vehicle, so as to realize the path tracking control of the vehicle.

[0063] ​In the fuzzy control, the input variable vehicle speed V is divided into 5 levels {very slow, slow, moderate, fast, very fast}, and the corresponding fuzzy sets are {LD, LX, M, RX, RD}; the lateral deviation EX between the actual vehicle trajectory and the target trajectory is divided into 5 levels {right large, right small, center, left small, left large}, and the corresponding fuzzy sets are {LD, LX, M, RX, RD}; the vehicle's forward distance L is d It is divided into 5 levels {closer, near, medium, far, farther}, and the fuzzy set is {VN, N, M, F, VF}.

[0064] Table 2 Fuzzy rules of fuzzy control

[0065]

[0066] The schematic diagram of the pure tracking model is as follows Figure 5 As shown, (x, y) is the next path point to be tracked, and the distance between the vehicle and the path target point is L d The angle between the current vehicle longitudinal body and the target path point is α. In the planned global path, in order to ensure that the rear axle of the vehicle passes through the target path point, the turning radius is obtained according to the triangle sine theorem;

[0067]

[0068] The foresight distance of pure tracking control includes the foresight distance output by fuzzy control and the minimum foresight distance L0. The minimum foresight distance calculated from the foresight distance output by fuzzy control is:

[0069] L0=L d -k v v

[0070] Where L0 is the minimum foresight distance, k v is the speed proportional coefficient, and v is the vehicle speed.

[0071] The expected front wheel turning angle calculated by the pure tracking model is:

[0072]

[0073] Among them, α(t) represents the angle between the vehicle body and the target path point per unit sampling time.

[0074] The effect of the unmanned vehicle path planning and tracking method described in the present invention is explained by establishing a controller through MATLAB / Simulink.

[0075] Figure 6 The curvature of the reference path at 4.8m, 17.2m, and 32.1m is 0.216m respectively. -1 、0.389m -1、0.667m -1 Taking a traditional pure tracking control look-ahead distance of L = 3m, simulations compared the pure tracking algorithm, the fuzzy pure tracking control algorithm, and the genetically optimized fuzzy tracking control algorithm. The optimal fuzzy tracking control curve exhibited a small turn at 31.6m, while the optimized curve avoided this. At 32.3m, where the curvature was greater, although both approaches deviated from the target path, the optimized fuzzy control approach produced a smaller deviation.

[0076] Figure 7 To optimize the foresight distance generated by the algorithm and achieve the best selection of the foresight distance; Figure 8 and Figure 9 Comparing vehicle lateral deviation and steering angle, the optimized fuzzy tracking control algorithm achieved a maximum lateral tracking deviation of 0.521m during cornering, a reduction of 44.9% and 14.3% compared to the pure tracking control algorithm and the fuzzy control algorithm. The optimized fuzzy tracking algorithm achieved a maximum front wheel steering angle of 8.061°, a reduction of 25.7% and 9.9%, respectively. In summary, the genetic optimization control strategy achieved superior control performance compared to both the traditional pure tracking control and fuzzy pure tracking control algorithms.

[0077] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the description and implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A method for tracking a path of an unmanned vehicle, characterized in that: include: Step 1: During the vehicle's driving process, a simplified kinematic model of the vehicle is established based on the set relationship of the global coordinate system; Step 2: Based on the genetic algorithm, the membership function of the fuzzy control is calculated with the vehicle longitudinal speed and lateral deviation as input and the vehicle forward sight distance as output; Step 3: Establish a fuzzy controller to obtain the vehicle's forward distance through the deviation between the vehicle's actual trajectory and the target trajectory and the vehicle's speed. Based on the tracking control method, the desired front wheel angle of the vehicle is calculated to complete the vehicle's path tracking control. The calculation formula for the expected front wheel turning angle of the vehicle is: Among them, α(t) represents the angle between the vehicle body and the target path point per unit sampling time.

2. The unmanned vehicle path tracking method according to claim 1, characterized in that: The simplified kinematic model of the vehicle is: Where, are the longitudinal velocity component, lateral velocity component and angular velocity component of the vehicle in the global coordinate system, v r is the vehicle's speed, is the heading angle, l is the wheelbase of the front and rear wheels, δ f is the turning angle of the front wheels.

3. The unmanned vehicle path tracking method according to claim 2, characterized in that: The maximum value of the objective function optimized by the genetic algorithm is: Where J(x) is the optimal function for evaluating the objective function controller; The calculation formula of the optimal function J(x) is: Where k is the duration, e(t) is the vehicle lateral deviation, ec(t) is the rate of change of the vehicle lateral deviation, u(t) is the control variable of the genetic algorithm, represents the membership function corresponding to the fuzzy control, and w1, w2, and w3 are the weighting coefficients of e(t), ec(t), and u(t), respectively.

4. The unmanned vehicle path tracking method according to claim 3, characterized in that: In the step 2, the membership function of the input speed obtained based on the genetic algorithm selects a combination of a trapezoidal membership function and a triangular membership function, the membership function of the input lateral deviation selects a combination of a Gaussian membership function and a triangular membership function, and the membership function of the output foresight distance selects a combination of a trapezoidal membership function and a triangular membership function.

5. The unmanned vehicle path tracking method according to claim 4, characterized in that: The domain of the vehicle speed V is [0, 2 m / s], the domain of the vehicle lateral deviation EX is [-0.6 m, 0.6 m], and the vehicle forward sight distance L d The domain of discourse is [1m,10m].

6. The unmanned vehicle path tracking method according to claim 5, characterized in that: The vehicle speed V is divided into 5 levels {very slow, slow, moderate, fast, very fast}, and the corresponding fuzzy sets are {LD, LX, M, RX, RD}; the lateral deviation EX of the actual vehicle trajectory from the target trajectory is divided into 5 levels {right large, right small, center, left small, left large}, and the corresponding fuzzy sets are {LD, LX, M, RX, RD}; the forward sight distance L of the vehicle is d It is divided into 5 levels {closer, near, medium, far, farther}, and the fuzzy set is {VN, N, M, F, VF}.

7. The unmanned vehicle path tracking method according to claim 5, characterized in that: The foresight distance in step 3 includes the foresight distance part and the minimum foresight distance part output by fuzzy control. The minimum foresight distance calculated based on the foresight distance output by fuzzy control is: L0=L d -k v in; Where L0 is the minimum foresight distance, k v is the speed proportional coefficient, and v is the vehicle speed.

Images