Collision-Free Navigation and Control for Autonomous Vehicle in Complex Urban Environments

This paper focused on the vehicle path planning in the highways and complex urban environments. At first, obstacles and road lines have been detected by sensors of the intelligent vehicle, thereupon the vehicle will be find the safe areas using the time distance method developed in this paper. Then, an appropriate path close to the intelligent decisions about human being would be chosen through the developed algorithm. There is the possibility of collision to surrounding vehicles in the areas where changing the lane is needed. Therefore, to prevent collision, a five orders polynomial curve is offered for each lane change maneuver. The reached maneuver is optimized based on the vehicle dynamic and allowed lateral acceleration. Finally, a suitable path to pass quite safely and without any collision through the obstacles is suggested. At the end, two main and different simulation scenarios included the lack of collision is verified by MATLAB software and the obtained path is controlled by the sliding mode controller. These simulations indicated effectiveness of this method. The lateral acceleration is obtained in allowed range for comfort of occupants in these scenarios.