Online incremental and heuristic path planning for Autonomous Ground Vehicle

Path planning for Autonomous Ground Vehicles( AGVs) in a dynamic environment is a quite complex and difficult task to perform when there is incomplete information about the surroundings. In dynamic environments, the system has to monitor changes about the kinematic behaviour of the obstacles regularly. The system has to replan its path to avoid the collision if necessary. This paper describes a novel algorithm to perform the above task in an effective manner. A 3D camera has been utilised as the exteroceptive sensor to form a simple network with the system. The sensed data is used to predict the trajectory of the moving obstacles using a spline function. Subsequently, the algorithm replans its path when the predicted obstacle tends to collide with the initial path. The proposed heuristic approach continuously updates obstacles and dynamically replans its predicted collision path. It is more effective in indoor environments, because it replans only the affected route which is necessary and avoids stopping during the vehicles traverse. The simulated results are presented to demonstrate the utility of this approach.