A Hybrid Artificial Potential Field:Genetic Algorithm Approach to Mobile Robot Path Planning in Dynamic Environments

摘要

In this paper, a hybrid Artificial Potential Field-Genetic Algorithm approach is developed and implemented for mobile robot path planning in dynamic environments. The hybrid approach first uses Grid Method where the mobile robot environment is represented by orderly numbered grids, each of which represents a location in the environment. Then, it applies Genetic Algorithm (GA), a global planner, to find an optimal path according to the current environment. The GA proposed here uses an evolutionary population initialization and genetic operators, which make the evolutionary process converge very efficiently. Finally, a new Artificial Potential Field method, a local planner, is applied to follow the path obtained by GA from one intermediate node to next intermediate node avoiding the obstacles. Experimental results clearly illustrate that the proposed hybrid approach works well on large scale dynamic environments.