Dissertation
Dissertation > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Robotics > Robot > Intelligent robots

Research on Strategy in the Middle-size League of RoboCup Based on Partheno-Genetic Algorithm

Author LiZuo
Tutor LiMaoJun
School Changsha University of Science and Technology
Course Control Theory and Control Engineering
Keywords RoboCup Strategy of Cooperation Formation Selecting Role Allocating Obstacle-Avoidance Partheno-Genetic Algorithm
CLC TP242.6
Type Master's thesis
Year 2011
Downloads 19
Quotes 0
Download Dissertation

System of soccer-robot is a system of multi-robot. The competition environment of the middle-size league of RoboCup is dynamic and with the characteristics of confrontation and uncertainty. So the appropriate designing of cooperative strategy is the key to its overall performance. Base on the research of the multi-robot cooperation, especially in the competition environment of the middle-size league of RoboCup, a multi-robot cooperation mechanism is designed based on partheno-genetic algorithm and combined with actual situation of the middle-size league of RoboCup. It achieves a dynamic cooperation with stability in multi-robot system. Especially for multi-robot role dynamic allocating problem and obstacle-avoidance strategy, an algorithm on role dynamic allocating and global obstacle-avoidance is designed, and ultimately a cooperation tactics of multi-robot is achieved.A dynamic formation selecting algorithm for the middle-size league of RoboCup is proposed. Multi agents are linked into a group to achieve a common goal through formation. A cooperation model is built to carry out the allocation and standing-location cooperation for a preassigned task. These strategies are performed to achieve dynamic cooperation based on the dynamic formation switching between different formations. Therefore, the requirement of different stages is satisfied, and the total ability of the team in the game is notably improved.In view of algorithm on role dynamic allocating of RoboCup, partheno-genetic algorithm is introduced to get a near-optimized solution of role allocating problem. Partheno-genetic algorithm, which is from the "survival of the fittest" mechanism of natural selection, gets the solution evolving into a near-optimized at last. And it solves the problem that the solution is easy to enter local optimization by using the greedy algorithm in the calculation. Experimental results indicate that partheno-genetic algorithm has the characteristic of global optimality, and it can satisfy the requirement of real-time. So the total ability of the team in the game is notably improved.Design of an obstacle-avoidance strategy based on partheno -genetic algorithms in the middle-size league of RoboCup which makes the robot avoid obstacles and find a shortest path from starting point to the target point, and the realizing process of this algorithm are introduced in detail. In this paper according to the characteristics of coding, a single point of gene clone operation is added in genetic manipulation of partheno-genetic algorithms which can accelerate the population evolutionary. The simulation results indicate that this algorithm can make the robot avoid obstacles and find the optimal path.This research for the coordination problem in the middle-size league of RoboCup combines both theory and reality. A coordination strategy of multi-robot in the real situations is designed, according to the existing platform. It is a foundation of further research and design of application on the coordination problem in the middle-size league of RoboCup.

Related Dissertations
More Dissertations