Research On Path Planning Of Mobile Robot Based On Genetic Algorithm

The mobile robot is an important branch of the field of robot. In particular, path-planning is critical to mobile robot system because it determines the quality of the robot's task. As a result, path-planning has attained more and more attention in the field of mobile robot.In this thesis mobile robot path planning includes two aspects: obstacle avoidance path planning and TSP path-planning problem. Obstacle avoidance path planning is achieving the best path from starting point to goal point by avoiding all barriers. The path searching depends on one or more optimization rules. TSP path-planning is the problem some known distance between the city, has a salesman to be visited these cities, and visit each city only once, finally return to the starting city, how to arrange the visit to the city of his order, make its total shortest length route of travel.This thesis first discusses the development situation of path planning in technology and application method, and points out the significance of this project and the main research contents.Then based on genetic algorithm and simulated annealing algorithm, analyzes their advantages and disadvantages. And the two algorithms combining constitute the genetic simulated annealing algorithm; it has strong global and local search ability, in large number of variables, especially prominent. The genetic simulated annealing algorithm is used to obstacle-avoiding path planning, and adopts new initial population generation algorithm, the simulation results show that this algorithm make mobile robot path planning to improve the obstacle-avoidance convergence speed, achieve good planning effect.Finally to discuss the use of genetic algorithm to solve TSP path-planning problem, the basic genetic algorithm to solve TSP path-planning problem has been improved. In order to solve the inconsistency between diversity and convergent speed, the paper also adopted probability greedy method produce original population, some greedy method to generate initial population, the initial population generation method slightly worse than the greedy method, but the level of individual diversity is better than greedy method. In the whole process of genetic algorithm, keep population diversity, improve convergence speed, the thesis use concept of similarity and classification into genetic algorithm, the higher level of individual uses hybrid algorithm to cross, and adopt the outstanding records mixed and"father and son mixed"strategy to ensure that the global convergence, and the simulation results prove the effectiveness of the algorithm.