Motion planning plays an essential role in designing self-driving functions for connected and autonomous vehicles. The methods need to provide a feasible trajectory for the vehicle to follow, fulfilling different requirements, such as safety, efficiency, and passenger comfort. In this area, algorithms must also meet strict real-time expectations, since, especially in an emergency, the decision time is limited, which raises a trade-off for the feasibility requirements. This article proposes a hierarchical path planning solution for evasive maneuvering, where a Twin Delayed DDPG reinforcement learning agent generates the parameters of a geometric path consisting of chlotoids and straight sections, and an underlying model predictive control loop fulfills the trajectory following tasks. The method is applied to the automotive double lane-change test, a common emergency situation, comparing its results with human drivers' performance using a dynamic simulation environment. Besides the test's standardized parameters, a broader range of topological layouts is chosen, both for the training and performance evaluation. The results show that the proposed method highly outperforms human drivers, especially in challenging situations, while meeting the computational requirements, as the pre-trained neural network and path generation algorithm can provide a solution in an instant, based on the experience gained during the training process.
Planning the optimal trajectory of emergency avoidance maneuvers for highly automated vehicles is a complex task with many challenges. The algorithm needs to decrease accident risk by reducing the severity and keeping the car in a controllable state. Optimal trajectory generation considering all aspects of vehicle and environment dynamics is numerically complex, especially if the object to be avoided is moving. This paper presents a hierarchical method for the avoidance of moving objects in an autonomous vehicle, where a reinforcement learning agent is responsible for local planning, while longitudinal and lateral control is performed by the low-level model-predictive controller and Stanley controllers. In the developed architecture, the agent is responsible for the optimization. It is trained in various scenarios to provide the necessary parameters for a polynomial-based path and a velocity profile in a neural network output. The vehicle performs only the first step of the trajectory, which is redesigned repeatedly by the planner based on the new state. In the training phase, the vehicle executes the entire trajectory via low-level controllers to determine the reward value, which realizes a prediction for the future. The agent receives feedback and can further improve its performance. Finally, the proposed framework was tested in a simulation environment and was also compared to human drivers’ abilities.
20 December 2024 in Engineering Applications of Artificial Intelligence