Research On Mobile Robot Path Planning Based On Genetic Algorithm

The research of mobile robot technology is the hot topic recently. Mobile robot path planning technology is one of the important domains in mobile robot technologies'research. Path planning is also the fundamental technique for mobile robot to complete other activities.The robot path planning is a kind of planning and offers the information of their work environment for robot. According to a certain performance indicators (such as distance, time, etc.), the robot path planning can search the optimal or sub-optimal path from the starting position to the target position. In this paper mobile robot path planning includes the content:The environment is static and the information of the environment is all known. Based on optimization criteria, to find an optimal or sub-optimal path in such an environment from the starting position to the target position. The path must avoid all the obstacles in the environment.Firstly this paper discusses the key technologies for existing mobile robot technology and the research situation at home and abroad, introduces the background of research areas of the subject and the significance of the subject, and lists several theories and algorithms that are commonly used in mobile robot path planning, analyzes the existing problems.Secondly the paper describes the mathematic basis of the simple genetic algorithm and sums up its desirable features and steps. Meanwhile, the paper shows the details of the composition of the simple genetic algorithm and analyzes the current commonly used genetic operators. And finally it gives a simple analysis of genetic algorithm convergence in different encoding mode.Finally the paper makes improvement on the basis of the simple genetic algorithm. The obstacles in the robot environment are described as polygons. The path is encoded into fixed-length decimal encoding. In the method, the fitness function contains explicit physic means and corresponding mutation operator is provided. It also applies the amendment operator and retains the best individuals, so that GA can be lead to an optimized result rapidly. Such an encoding mechanism overcomes the shortcomings of variable-length encoding mechanism and other fixed-length binary encoding mechanisms where special decoding and genetic operators are needed, resulting in simplicity and effectiveness of the algorithm. Improved crossover operator ensures diversity of population, it is also conducive to convergence of the algorithm. The computer simulation experiment shows that, the method has good path planning ability. Path planning algorithm based on improved GA is implemented in the AS-MF09 mobile robot for the fire-fighting. The result of the experiments demonstrates the proposed method's validity.