A GENETIC NAVIGATION ALGORITHM FOR AUTONOMOUS MOBILE ROBOTS

摘要

The use of genetic algorithms GAs for solving the navigation problem of mobile robots is explored in this paper. Without using any logical model of the robot's world, the proposed approach evolves feasible robot paths by performing an adaptive search on populations of candidate robot actions. Unlike standard GAs, the proposed approach uses variable length chromosome structures and a set of special genetic operators. More importantly, the genetic navigation algorithm GNA can adapt its behavior for dealing with different types of static or changing simulated environments. The performance of the algorithm is examined through a plethora of computersimulated robot navigation experiments carried out on various types of planar grid environments. Furthermore, experimental comparisons between the GNA and two other heuristics are provided showing a substantially better performance for the GNA.