Optimal path planning of autonomous navigation in outdoor environment via heuristic technique

Abstract

Path planning for automated guided vehicles (AGVs) operating in an outdoor environment is a key task in logistic applications. This paper highlights a new approach for finding collision-free path for AGVs in an outdoor environment using satellite images. Satellite images are transformed into occupancy grids that divide the environment, obstacles (building, trees and pedestrian path etc.,) as black grids and free routes as white grids in order to find an obstacle free path. The proposed navigation control strategy is to localise the AGVs vehicle route until any approaching obstacle is detected and to turn to the safest area before continuing its path. The navigational model was developed in the MATLAB/Simulink 2D virtual environment to visualise the path navigation. Particle swarm optimization and its five variants were used to obtain the obstacle free path. Experiments were conducted in various environments to verify the efficiency and performance of the algorithm. Validation with a bench mark instance ensures that the proposed algorithm found a shorter path with less computation time.

Keywords

• Localization
• Satellite image
• Path planning
• Particle swarm optimization