CN-121994243-A - Manual potential field method vehicle obstacle avoidance control method based on PSO optimized potential field parameters

CN121994243ACN 121994243 ACN121994243 ACN 121994243ACN-121994243-A

Abstract

A manual potential field method vehicle obstacle avoidance control method based on PSO optimized potential field parameters belongs to the technical field of automatic driving and intelligent vehicle motion planning. Aiming at the problems of the traditional artificial potential field method in the aspects of local minimum value, parameter sensitivity, path oscillation and the like, the invention provides a real-time obstacle avoidance strategy combining a particle swarm optimization algorithm with the artificial potential field method. According to the method, five key parameters including an attraction gain coefficient, an obstacle repulsive force gain coefficient, a boundary repulsive force gain coefficient, a speed potential field repulsive force gain coefficient and an obstacle influence distance threshold value in an artificial potential field are optimized on line through a particle swarm, and an adaptability function comprehensively considering path length, smoothness and safety is designed. The method can effectively avoid local minimum points, remarkably improve the path smoothness and the running stability, simultaneously realize energy conservation by more than 5% in various dynamic scenes, and has better environmental adaptability and engineering practicability.

Inventors

ZHAO JINGHUA
Yang Peici
LI XIAONA
LI XIAOKUI
YUAN HE

Assignees

吉林师范大学

Dates

Publication Date: 20260508
Application Date: 20260209

Claims (1)

1. A manual potential field method vehicle obstacle avoidance control method based on PSO optimized potential field parameters is characterized by comprising the following steps: S1, building a vehicle motion environment model fusing multiple potential fields: S11, target gravitation potential field: (1) in the formula, Gravitational gain coefficient; representing the current position of a vehicle To sub-target Euclidean distance of (c); To point to sub-targets A unit direction vector; S12, an obstacle repulsive force potential field: (2) in the formula, Gain coefficient for the obstacle repulsive force; a minimum distance from the vehicle to the obstacle; affecting a distance threshold for the obstacle; a unit direction vector directed from the obstacle to the vehicle; is a target regulatory factor; and (3) with Euclidean distances respectively for current target adjustment and previous target adjustment ; To point to sub-targets A unit direction vector; S13, road boundary repulsive force potential field: (3) in the formula, Gain coefficient for boundary repulsive force; For vehicle position To the nearest boundary Is a distance of (2); The corresponding boundary repulsive force is the negative gradient of the potential field: (4) in the formula, A unit direction vector directed from the boundary to the vehicle; representing a road boundary repulsive potential field function In the vehicle position Gradient vector at; s14, repulsive force of a speed potential field: (5) in the formula, Gain coefficient for repulsive force of velocity potential field; Is a velocity vector of the vehicle relative to the dynamic obstacle; the corresponding velocity potential field repulsive force is: (6) in the formula, Representing a velocity repulsive potential field function In the vehicle position Gradient vector at; s15, resultant force applied to vehicle Vector sum of all attractive and repulsive forces: (7) in the formula, The number of the obstacles detected at the current moment is the number of the obstacles; Is the first Repulsive force generated by the obstacle; S2, constructing PSO-APF fusion optimization strategy Five key parameters in APF are encoded as position vectors of PSO particles: (8) in the formula, As the coefficient of the gain of the attraction force, For the gain factor of the obstacle repulsive force, Is the gain coefficient of the boundary repulsive force, As the gain factor of the repulsive force of the velocity potential field, Affecting a distance threshold for the obstacle; the position of each particle represents a set of APF parameters whose velocity and position update formula is as follows: (9) (10) in the formula, As a vector of the velocity of the particles, As a vector of the position of the particles, As the weight of the inertia is given, , In order for the learning factor to be a function of, , A uniform random number within the range of [0,1 ]; Is a particle At the iteration number Velocity vector at that time; Is a particle At the iteration number Updated velocity vector at that time; Is a particle At the iteration number A position vector at that time; Is a particle At the iteration number A position vector at that time; Is a particle Is a single historical optimal position vector; A group global optimal position vector; designing fitness functions To comprehensively evaluate parameter performance, its function fuses path length Smoothness degree Safety of Failure penalty : (11) Wherein, the Is a path length weight coefficient; Is a smoothness weight coefficient; is a safety weight coefficient; punishment of weight coefficients for failure; s3, realizing global-local rolling optimization planning framework 1) The global path planning stage is that an A-algorithm is adopted, and an optimal reference path from a starting point to an end point is generated based on environment priori information; 2) A PSO online parameter optimization stage, namely starting a PSO optimizer in a local rolling window by taking a global path as a reference; 3) And in the local obstacle avoidance execution stage, initializing an APF model by adopting the optimized parameters.

Description

Manual potential field method vehicle obstacle avoidance control method based on PSO optimized potential field parameters Technical Field The invention belongs to the technical field of automatic driving and intelligent vehicle motion planning. Background In an automatic driving system, real-time and reliable local path planning and dynamic obstacle avoidance capability are cores for realizing safe autonomous navigation. The artificial potential field method is widely applied to real-time obstacle avoidance planning due to the advantages of simple model, high calculation efficiency, easiness in implementation and the like. The method is to construct a virtual potential field to lead the vehicle to move under the guidance of the resultant force of the attraction of the target point and the repulsive force of the obstacle. However, the traditional artificial potential field method has obvious limitations that 1) the parameter sensitivity is strong, key parameters such as potential field gain coefficient and the like are usually set depending on experience, fixed parameters are difficult to adapt to complex and changeable dynamic environments, and path oscillation and planning failure are easy to cause, 2) the problem of local minima is that a vehicle is easy to sink into a local minima with balanced attraction and repulsion so as not to reach a target, and 3) the problem that the target is unreachable is that when an obstacle approaches to the target point, the excessive repulsion can prevent the vehicle from reaching the target. To overcome the above problems, researchers have proposed various improvements such as introducing virtual target points, combining velocity potential fields, improving repulsive force functions, and the like. Although the methods have a certain effect, the performance of the methods still depends on parameter configuration seriously, and the methods lack self-adaptive capability to dynamic changes of environment. As a high-efficiency intelligent optimization algorithm for the group, the particle swarm optimization algorithm has the characteristics of few parameters, rapid convergence, suitability for online optimization and the like, and provides a new idea for solving the problem of adaptive adjustment of APF parameters. However, the existing researches focus on theoretical analysis and simulation verification, PSO and APF are deeply fused, and real-time and robust online parameter optimization and obstacle avoidance control are realized in a dynamic environment under a limited perception condition, so that deep exploration is still needed. Particularly in specific scenes such as industrial parks and closed parks, a vehicle sensing system is often limited by low-cost laser radars, the problems of limited sensing range, lack of semantic information and the like exist, and higher requirements are put on the environmental adaptability and reliability of obstacle avoidance algorithms. Disclosure of Invention The invention aims to realize on-line self-adaptive optimization of APF key parameters, thereby effectively overcoming local minimum values, improving path smoothness and running stability and realizing an artificial potential field method vehicle obstacle avoidance control method based on PSO optimized potential field parameters for energy-saving driving under the limited perception condition. The method comprises the following steps: S1, building a vehicle motion environment model fusing multiple potential fields: S11, target gravitation potential field: (1) in the formula, Gravitational gain coefficient; representing the current position of a vehicle To sub-targetEuclidean distance of (c); To point to sub-targets A unit direction vector; S12, an obstacle repulsive force potential field: (2) in the formula, Gain coefficient for the obstacle repulsive force; a minimum distance from the vehicle to the obstacle; affecting a distance threshold for the obstacle; a unit direction vector directed from the obstacle to the vehicle; is a target regulatory factor; and (3) with Euclidean distances respectively for current target adjustment and previous target adjustment;To point to sub-targetsA unit direction vector; S13, road boundary repulsive force potential field: (3) in the formula, Gain coefficient for boundary repulsive force; For vehicle position To the nearest boundaryIs a distance of (2); The corresponding boundary repulsive force is the negative gradient of the potential field: (4) in the formula, A unit direction vector directed from the boundary to the vehicle; representing a road boundary repulsive potential field function In the vehicle positionGradient vector at; s14, repulsive force of a speed potential field: (5) in the formula, Gain coefficient for repulsive force of velocity potential field; Is a velocity vector of the vehicle relative to the dynamic obstacle; the corresponding velocity potential field repulsive force is: (6) in the formula, Representing a velocity re