An automatic navigation system for vision guided vehicles using a double heuristic and a finite state machine

A navigation system for automatic vision-guided vehicles which uses an efficient double heuristic search algorithm for path planning is presented. It is capable of avoiding unknown obstacles and recovering from unidentifiable locations. A linked list representation of the path network database makes the implementation feasible in any high-level language and renders it suitable for real-time application. Extensive simulated experiments have been conducted to verify the validity of the proposed algorithms. The combination of the techniques of robot navigation in unexplored terrain and the global map method proved to be a valid technique for automated guided vehicle (AGV) guidance. A learning mechanism is used in the AGV by updating the path network during navigation. Simulated results supported all the theoretically expected conclusions, since the robot planned its path correctly between the requested nodes and maneuvered its way around the obstacles. Overall, the results were very encouraging