Genetic Path Planning Method For Mobile Robot In Uncertain Environments

In the field of route planning , genetic programming has been applied to the simulation of Mobile Robot Wall-Following , artificial ant and other problems successfully and shown a satisfying result .But , many problems remain unsolved: redundant actions may arise while robot perform some special tasks ,the local convergence that may arise in the course of evolution confine the appliance of GP in on-line route planning ,and there is no effective way that can be used to forecast the adaptation of individual program. This paper does research into these problems and presents some solutions changing the terminal sets and designing new fitness function are used to reduce redundant actions and changing mutation probability in the run is used to reduce the probability of local convergence and changing the evolution environment in the run is used to improve the adaptation capability of individual program and adaptation capability of individual program can be forecasted by studying the current environment .In addition, this paper also does research into environmental model according to the algorithm so that the advantages of GP can be brought into play .The implementation of the genetic programming project was written in C++ .The result shows that these solutions work well.