Research On Mobile Robot Path Planning Based On Intelligent Computing

Path planning is an important field of the mobile robot research, and as a basic part of the mobile robot navigation, it aims to search for a non-collision path from the start to the target according to a performance indicator. Intelligent computing, arousing extensive attentions as a special way of solving non-linear problems, is a calculating method of gaining or expressing knowledge, and a subject of simulating the intelligence. Researching the path planning of mobile robot with intelligent means is the main task of this dissertation. They are as follows:1 Introducing the definitions and classifications of the robots in detail; discussing the signification, developments home and abroad, main research areas of the mobile robots technologies; explaining the development situation of series intelligent calculations: neural network, genetic algorithm, simulating anneal and fuzzy control synoptically.2 Introducing the hardware system, software system and motion model of the Pioneer2 robot.3 Proposing a path planning algorithm of mobile robot based on the neural network, studying path planning with the principle of the neural network, i.e. constructing a neural network model to describe the robot's environment, and then searching for the non-collisions path.4 Presenting a simulating anneal algorithm based on the NN path planning, i.e. using the simulating anneal algorithm to solve the local minimum problem of the NN path planning algorithm.5 Employing the genetic algorithm to improve the NN path planning, i.e. establishing a neural network model to describe the robot's environment, and then searching for a optimal path without touch with the genetic algorithm.6 Describing a path planning algorithm based on fuzzy neural network in order to improve the mobile robot performances in the dynamic environment, which can do self-learning on line. The results of simulation have verified that the method is effective.