Mobile Robot Global Path Planning Based On Immune Genetic Algorithm

As the rapid development of the computer, internet, information, automation and artificial intelligence, a new stage comes from the research of mobile robot. As the same time, the need of military progress, the development and application of space resource and ocean resource provide wide space room for the development of mobile robot.In this paper, the history, classification, definition and technology of robot are introduced, and some issues and methods of global path planning were narrated, also the advantages and shorts of genetic algorithm, immunity algorithm and other algorithms were introduced. They are the base of the paper' theoretical research.A method of global path planning based-on immunity-genetic algorithm was provided in this paper. First the neural network model of environmental information was constructed in the working space for robot, also the relationship between the collision-free path and the neural network output was established based-on this model, then the collision-free demand and the path optimization demand were fused to a simple fitness function for immune genetic algorithm. Antibody selective probability was expressed as a fusion function based-on antibody vector distance and antibody density, the antibody diversity and maturation convergence were ensured synchronously. Finally the simulation results show that the performance was improved by contrast with genetic algorithm, also simulation results demonstrate that the algorithm for global path planning was feasible and valid.At last, a synthetic experimentation was done about robot path planning. Based-on the immunity-genetic algorithm, a path was obtained to let the robot round the barrier first, then the robot used the self-sensor to estimate the surroundings information and identify letters, finally it reached the destination. From the results of experiment, the immunity-genetic algorithm designed in this paper was validity and availability. Meanwhile, it also demonstrates the method of identify letters based on the characteristics feature of first order histogram.