Vehicle maneuver evaluation in emergency condition

Abstract

Emergency vehicles getting stuck in traffic jams have always been a great concern in the cities. This would be more concern when the car is autonomous. So, evaluating the vehicle's maneuver in the presence of an emergency vehicle is required. This problem has been questioned in different ways. This paper attempts to deal with the performance of the vehicle in front of the emergency vehicle by considering the emergency vehicle's right-of-way. To evaluate the behavior of such a vehicle, we propose an algorithm that comprises three parts. In the first part, using a decision tree model, all feasible maneuvers that can be done by the front vehicle in the presence of an emergency vehicle are predicted. In the next part, the performed maneuver is detected using Hidden Markov Model. Finally, the best possible maneuver is compared with the performed maneuver. The proposed algorithm is implemented on a simulator, also developed in this research. The simulator generates different driving behaviors to train the models and evaluate the proposed algorithm.