Algorithm Research Of Dynamic And Static Path Planning For Soccer Robots

Soccer robots system is a complex artificial intelligent system that combines machinery and electronics, sensor information fusion, intelligent control theory, communication, computer control and more other subjects, and for this reason, it provides a good platform for the research of robotics and artificial intelligence. Robot soccer competition is a complex dynamic environment which is real-time, confrontational, and uncertain, there are plenty of technologies, such as multi agent cooperation, strategy acquisition, real-time reasoning, robotics, perception of information fusion technology and so on, these above are all integrated within the competition, so, on this platform for the study of robot path planning is a very challenging topic. Path planning is an important part of soccer robot decision-making system, it is to find an optimal path as the goal, and makes choices through a variety of complex algorithm for robots soccer confrontation strategy and cooperation strategy.This topic mainly puts robot soccer competition as the background, using artificial potential field method and the improved particle swarm intelligence algorithm to complete study of the soccer robot’s path planning problem in obstacle avoidance, dynamic interception and goal pursuit.Traditional artificial potential field method is used in the soccer robot’s path planning, it analyzes the unreachable destination problem and the partial minimum value problem that artificial potential field method shows. The robot cannot reach the target when there is obstacle near the target, and to solve this problem, the relative distance potential function method is proposed, by bringing in the distance parameters of the robots and the target into the repulsive potential function to form a new repulsive force function, the unreachable destination problem can be overcome. To solve the problem of partial minimum value, the method of artificial potential field based on deflection angle has been put forward. This method shows that when the resultant force of the robot is zero, by changing the repulsion force component direction and the direction of the deflection force method can make it out of the problem of partial minimum value. To solve the problem of un-optimal path curve under general condition, we put forward the optimization method that according to the judgment direction of obstacle avoidance obstacle position. In order to adapt to the soccer robot confrontational competition environment,we put forward relative speed dynamic artificial potential field method, and the dynamic obstacle avoidance of robot, and target simulation, path smooth safety.By using particle swarm algorithm in soccer robot’s path planning, the problem of robot soccer environment modeling’s complexity and the problem of algorithm’s limitations can be solved, at the meantime, by using the polar coordinate model, the improved particle swarm algorithm of path planning method is proposed. With the advantages of polar coordinates’ easy calculation and the path’s clear simulation, the environmental modeling puts the soccer robot path codes in the polar coordinates of two-dimensional, then judging the distance between robot and obstacles and between robot and destination to complete the path planning. In order to make the optimal algorithm and obstacle avoidance’s effect obvious, the fitness function and the degree-of-safety constraint function are applied to choose the best planning path. By introducing nonlinear dynamic decreasing inertia weight strategy to improve the algorithm, and using Ackley function test to verify the convergence of the algorithm. Finally under the dynamic environment, the robot obstacle avoidance, chase and antagonistic action simulation analysis are made, the result shows that the planned path has the advantages of obvious ability of obstacle avoidance and the shortest path.MATLAB is used to simulate and analyze. We simulate and analyze the soccer robot’s contrarious practice under the condition of 1vs2 by using improved artificial potential field method and particle swarm optimization(PSO) algorithm, and the result proves the validity and practicability of the two algorithms, besides, proves that the particle swarm algorithm in the aspect of convergence, accuracy and shortest path is better than those of artificial potential field method.