Urban scene-oriented automatic driving variable neighborhood searching and trajectory planning method

Abstract

The invention discloses a city scene-oriented automatic driving variable neighborhood searching and trajectory planning method, which relates to the technical field of automatic driving control, and comprises the following steps of: positioning a vehicle according to surrounding environment perception and a high-precision positioning device; performing mathematical description on rectangular and trapezoidal neighborhoods; performing trajectory planning and optimization on the search of the neighborhood of the independent interval; ensuring the continuity and stability of the track under the variable neighborhood search by using an average residence time method; on the premise of variable neighborhood search, according to traffic signs in an urban complex road scene, in combination with dynamic constraints of vehicles, trajectory planning of vehicle driving is carried out, and trajectory planning is divided into planning in a single neighborhood and complete trajectory planning in the urban complex road scene. The stability of switching of different tracks is ensured in complete track planning, so that the differential information capturing capability and the path planning and control capability of the intelligent vehicle in a complex scene of an urban intersection are improved, and the driving safety of the intelligent vehicle is improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving control technology, and in particular to an autonomous driving variable neighborhood search and trajectory planning method for urban scenarios. Background Art

[0002] With the acceleration of urbanization, urban traffic management faces unprecedented challenges. The complexity of urban traffic networks continues to increase, traffic flows are huge and the patterns are varied, and the interweaving of pedestrians, vehicles and non-motor vehicles makes traffic management extremely complicated. This complexity not only affects travel efficiency, but also increases the risk of traffic accidents, posing a threat to the property and life safety of urban residents. The Ministry of Transport's implementation of the "Urban Public Transportation Regulations" aims to improve the level of urban traffic management through intelligent means. In view of the special layout and traffic flow of urban scenes, the present invention invents an automatic driving variable neighborhood search and trajectory planning method for urban scenes to solve the current problems of insufficient capture of differential information, poor path planning and control, and high accident risks faced by smart cars in complex scenes such as urban intersections. Summary of the invention

[0003] The purpose of the present invention is to provide an urban scene-oriented automatic driving variable neighborhood search and trajectory planning method to solve the problems existing in the prior art mentioned in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A variable neighborhood search and trajectory planning method for autonomous driving in urban scenarios, comprising the following steps:

[0006] S1: Position the vehicle based on surrounding environment perception and high-precision positioning device;

[0007] S2: Mathematical description of rectangular and trapezoidal neighborhoods;

[0008] S3: trajectory planning and optimization for the search of the neighborhood of the independent interval;

[0009] S4: The average dwell time method is used to ensure the continuity and stability of the trajectory under variable neighborhood search.

[0010] Preferably, in S1, positioning the vehicle includes dividing various positions of the vehicle during driving in a complex urban scene, and confirming a triggering vehicle neighborhood search transition point.

[0011] Preferably, in S2, the mathematical description of the neighborhood facilitates the division of the vehicle's surrounding environment, and the complete trajectory is planned by superimposing the divisions of different neighborhoods to form a complete route.

[0012] Preferably, in S3, trajectory planning and optimization for searching the neighborhood of the independent interval include planning within a rectangular search neighborhood and planning within a trapezoidal search neighborhood. In each interval of the neighborhood switching search, the surrounding environment is segmented and superimposed to obtain regular shapes such as rectangles and trapezoids, thereby planning a complete trajectory.

[0013] Preferably, in S4, ensuring the stability and continuity of the complete trajectory is divided into the following steps:

[0014] S41: Using Lyapunov functions to determine the stability of trajectories within a single neighborhood and their convergence time;

[0015] S42: Use the average dwell time method to ensure the stability of neighborhood search switching.

[0016] Preferably, the mathematical description of the trapezoidal search neighborhood is:

[0017] The mathematical expression of the trapezoidal search neighborhood is O(x0, y0) to represent the initial position of the smart car. The lengths of the sides l1 and l2 of the neighborhood are both r, the top side is l3, the height is s, and the length of the bottom side is d, as shown in the following formula:

[0018]

[0019] Trajectory planning within the trapezoidal neighborhood: Determine the starting point and end point within the trapezoidal search neighborhood, use the quintic interpolation polynomial to plan within the neighborhood, obtain the trajectory candidate set, and then obtain the optimal trajectory curve.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] The present invention plans the vehicle's driving trajectory based on variable neighborhood search, traffic signs in complex urban road scenes, and vehicle dynamic constraints. The trajectory planning is divided into planning within a single neighborhood and complete trajectory planning in complex urban road scenes. The stability of switching between different trajectories is guaranteed in the complete trajectory planning, thereby improving the smart car's ability to capture differential information and plan and control paths in complex urban intersections, thereby improving the driving safety of the smart car. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a schematic diagram of the location and division of the urban intersection environment according to the present invention.

[0023] Figure 2 The vehicle body coordinate system diagram established for the present invention. DETAILED DESCRIPTION

[0024] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0025] See also Figure 1-2 , the present invention provides the following technical solutions:

[0026] A variable neighborhood search and trajectory planning method for autonomous driving in urban scenarios, comprising the following steps:

[0027] S1: Positioning the vehicle based on surrounding environment perception and high-precision positioning devices; positioning the vehicle includes dividing the various positions of the vehicle during driving in complex urban scenes, and confirming the triggering point of the vehicle neighborhood search.

[0028] Among them, the location division includes: the straight section before entering the urban intersection, the waiting area at the urban intersection, the central driving area of ​​the urban intersection, and the exit area of ​​the urban intersection.

[0029] Waiting area at urban intersections: When a vehicle is about to enter a city intersection, the area where the vehicle travels in a straight line at the stop line of the zebra crossing at the city intersection is called the waiting area at the urban intersection.

[0030] Central driving area of ​​urban intersections: The area from where vehicles leave the stop line and enter the core area of ​​a city intersection according to traffic lights to the stop line where they leave the city intersection is called the central driving area of ​​a city intersection.

[0031] Urban intersection departure area: The driving range from the stop line of the urban intersection to the departure of the located urban intersection scene is called the urban intersection departure area.

[0032] After locating and dividing the urban intersection environment, the stop line at the urban intersection is used as the switching point for the trapezoidal neighborhood search. When the vehicle leaves the urban intersection scene, the stop line is also used as the switching point for the rectangular neighborhood search. Figure 1 As shown in the figure, the black dotted line in the intersection is the stop line for vehicles, which corresponds to the zebra crossing of the urban intersection. That is, the vehicle needs to switch the neighborhood search twice when passing through the urban intersection. The first time is to switch from the rectangular neighborhood search to the trapezoidal neighborhood search, and the second time is to switch from the trapezoidal neighborhood search to the rectangular neighborhood search.

[0033] S2: Mathematical description of rectangular and trapezoidal neighborhoods; the mathematical description of the neighborhood facilitates the division of the vehicle's surrounding environment. The complete trajectory is planned by dividing and superimposing different neighborhoods. The mathematical description of the neighborhood facilitates the division of the environment around the vehicle and the superposition of neighborhoods to form a single neighborhood, realizing the planning of the trajectory within an independent neighborhood.

[0034] As attached Figure 2 As shown in the figure, in order to combine the vehicle's driving and the description of the rectangular neighborhood and trapezoidal neighborhood search, it is necessary to establish a vehicle coordinate system, with the front axle of the car as the x-axis and the forward direction of the car as the y-axis. The mathematical description of the neighborhood is closely related to the superposition of the trajectory search within the neighborhood. The specific mathematical description is as follows:

[0035] Take the trapezoidal search neighborhood as an example for mathematical description:

[0036] The mathematical expression of the trapezoidal search neighborhood is O(x0, y0) to represent the initial position of the smart car. The lengths of the sides l1 and l2 of the neighborhood are both r, the top side is l3, the height is s, and the length of the bottom side is d, as shown in the following formula:

[0037]

[0038] Trajectory planning within the trapezoidal neighborhood: Determine the starting point and end point within the trapezoidal search neighborhood, use the quintic interpolation polynomial to plan within the neighborhood, obtain the trajectory candidate set, and then obtain the optimal trajectory curve.

[0039] S3: Trajectory planning and optimization for the search of the neighborhood of the independent interval; Trajectory planning and optimization for the search of the neighborhood of the independent interval includes planning within the rectangular search neighborhood and planning within the trapezoidal search neighborhood. In each interval of the neighborhood switching search, the surrounding environment is segmented and superimposed to obtain regular shapes such as rectangles and trapezoids, thereby planning a complete trajectory.

[0040] The trajectory planning of neighborhood search within an independent interval first determines the size of the neighborhood of a single interval, and then determines the starting and ending points of the independent interval based on the vehicle size. The front axle center of a conventional vehicle is used as the starting point, and the starting point of the next interval is used as the end point. The driving route within the neighborhood of a single interval is planned based on the quintic interpolation polynomial.

[0041] Firstly design the sliding surface function S to track and control the neighborhood search trajectory, and then design the Lyapunov function to judge the stability of the interval neighborhood search system and the convergence time t of the neighborhood. s .

[0042] According to the defined yaw angle tracking error and its changing rate, the sliding membrane surface function S is designed.

[0043]

[0044] In formula 1, ε i is the yaw angle tracking error, is the yaw angle, is the expected yaw angle;

[0045]

[0046] In formula 3, β>0, q1, q2 are positive odd numbers, and 1<q1 / q2<2; the control rate U can be obtained by differentiating the sliding membrane surface function.

[0047] The Lyapunov function V1 needs to be defined:

[0048]

[0049] It can prove the convergence of the control system and the convergence time of the system on the sliding membrane surface;

[0050] Redefine Lyapunov function V2;

[0051]

[0052] In formula 5, e i is the comprehensive error. After taking its derivative and integrating it, we can get the convergence time t s have to;

[0053]

[0054] In formula 6 α>0, v1(0) is the initial energy of the system, and p1 and p2 are positive odd numbers.

[0055] S4: Use the average dwell time method to ensure the continuity and stability of the trajectory under variable neighborhood search; To ensure the stability and continuity of the complete trajectory, it is divided into the following steps:

[0056] S41: Using Lyapunov functions to determine the stability of trajectories within a single neighborhood and their convergence time;

[0057] S42: Use the average dwell time method to ensure the stability of neighborhood search switching.

[0058] The search of a single neighborhood will not fall into the local optimum, ensuring the non-convexity of the trajectory; the time interval between two neighborhood search switches when the vehicle is driving is greater than the average residence time, ensuring the stability of the entire system when the vehicle is driving at urban intersections.

[0059] The core technology of solving variable neighborhood search trajectory lies in how to solve the discontinuity when switching between different neighborhoods and the search of the next neighborhood falling into local optimality when the neighborhood is switched; in view of the discontinuity when switching between different neighborhoods, the average residence time method in the switching system is introduced to ensure that the interval between any two consecutive switches is greater than the average residence time, which can ensure the continuity and exponential stability of the switching system. Define N σ (t, τ) is the number of switching times within the period 0≤τ≤t, Save(τ D , N0) are all switches that satisfy the following inequality:

[0060]

[0061] where τ D is the average dwell time, which refers to the average residence time of the system under a neighborhood search, and N0 is the amplitude of the switching signal; as long as τ D If it is large enough, any number of switches in this set can ensure the exponential stability of the switching system. The number of switches at urban intersections is generally only two. In special cases, the stability of the switching trajectory can be guaranteed by satisfying the above conditions.

[0062] When the search for the next neighborhood falls into a local optimum when switching the neighborhood search, according to the above step S3, the convergence time t of the neighborhood search in a single interval is calculated. s , when the vehicle gives a switching signal, the dwell time τ is guaranteed D Greater than the convergence time, i.e. τ D ≥t s This ensures that when switching neighborhood search, the next neighborhood search interval will not fall into the local optimum.

[0063] The interval between two consecutive neighborhood switches is t0, and the convergence time of neighborhood search in a single interval is t s , average residence time τ D The interval time is dynamically adjusted according to the vehicle driving conditions in the urban intersection and simulation experiment; ensuring that: t s ≤τ D ≤t0, that is, the vehicle can ensure the continuity and stability of the planned trajectory in the urban intersection scenario.

[0064] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A variable neighborhood search and trajectory planning method for autonomous driving in urban scenarios, characterized in that: The following steps are involved: S1: Position the vehicle based on surrounding environment perception and high-precision positioning device; S2: Mathematical description of rectangular and trapezoidal neighborhoods; S3: trajectory planning and optimization for the search of the neighborhood of the independent interval; S4: The average dwell time method is used to ensure the continuity and stability of the trajectory under variable neighborhood search.

2. The method for variable neighborhood search and trajectory planning for urban scenarios of autonomous driving according to claim 1, characterized in that: In S1, positioning the vehicle includes dividing the various positions of the vehicle during its driving in a complex urban scene, and confirming a triggering point for the vehicle neighborhood search.

3. The method for variable neighborhood search and trajectory planning for urban scenarios of autonomous driving according to claim 1, characterized in that: In S2, the mathematical description of the neighborhood facilitates the division of the vehicle's surrounding environment, and the complete trajectory is planned by superimposing the divisions of different neighborhoods.

4. The method for variable neighborhood search and trajectory planning for urban scenarios of autonomous driving according to claim 1, characterized in that: In S3, trajectory planning and optimization of the search of the independent interval neighborhood include planning within the rectangular search neighborhood and planning within the trapezoidal search neighborhood. In each interval of the neighborhood switching search, the surrounding environment is segmented and superimposed to obtain regular shapes such as rectangles and trapezoids, thereby planning a complete trajectory.

5. The method for variable neighborhood search and trajectory planning for urban scenes of autonomous driving according to claim 1, characterized in that: In S4, the stability and continuity of the complete trajectory are ensured, which is divided into the following steps: S41: Using Lyapunov functions to determine the stability of trajectories within a single neighborhood and their convergence time; S42: Use the average dwell time method to ensure the stability of neighborhood search switching.

6. The method for variable neighborhood search and trajectory planning for urban scenes of autonomous driving according to claim 4, characterized in that: The mathematical description of the trapezoidal search neighborhood is: The mathematical expression of the trapezoidal search neighborhood is O(x0, y0) to represent the initial position of the smart car. The lengths of the sides l1 and l2 of the neighborhood are both r, the top side is l3, the height is s, and the length of the bottom side is d, as shown in the following formula: Trajectory planning within the trapezoidal neighborhood: Determine the starting point and end point within the trapezoidal search neighborhood, use the quintic interpolation polynomial to plan within the neighborhood, obtain the trajectory candidate set, and then obtain the optimal trajectory curve.

Images