A navigation strategy based on global geographical planning and local feature positioning for mobile robot in large unknown environment

A novel navigation strategy for mobile robot in large unknown environment is put forward here. Globally, the robot´s path is planned with genetic algorithm, which is better than A* algorithm and neural networks. Locally, the robot avoids obstacle and moves with some basic behaviors and path memory. With GPS and INS information fusion, the robot can go from one sub target to another sub target with enough accurate positioning. Finally, the robot reaches the target accurately with feature positioning. Finally experiments are carried out on our mobile robot in many kinds of man-made complex environments, and the navigation strategy based on global geographical planning and local feature positioning is proved successful, which give a solid support for our further research on mobile robot navigation in unknown environments