2 October 2020 in Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering
There are two shortcomings in the application of traditional A* algorithm in the path planning of autonomous driving. One is that the vehicle environment description method suitable for the A* algorithm is not given; the other is that the vehicle contours and kinematic constraints are not considered. Therefore, according to the characteristics of unstructured environment, this paper presents an environment description method combining global navigation layer and local planning layer, and proposes a local motion planning algorithm based on the improved A* algorithm for autonomous driving vehicles in unstructured environment. In the improved algorithm, profile collision is avoided by setting redundant security space, and the cost of path curvature is considered in heuristic function design. Compared with the original algorithm, it can improve the smoothness of the path, so as to get a path more satisfied with vehicle motion constraints. Simulation results show that the improved algorithm can avoid vehicle contours collision and output a smoother path.
25 September 2021 in Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering
Automated driving motion planning algorithms require a high degree of real-time performance. In a closed structured road environment, making full use of the road information helps to reduce the computational effort and improve the real-time performance of the algorithm. Therefore, this paper proposes a real-time automated driving motion planning process suitable for closed structured road environments. Based on the environmental information of the structured road, the key factors affecting the motion decision are extracted. The scenarios are classified according to the key factors, and decision rules are established for the vehicle based on the headway. In addition, based on a pure tracking algorithm, a motion planning method based on dynamic target points is proposed, which takes into account both single-cycle trajectory planning and vehicle motion continuity. Finally, a real-vehicle test is conducted under two typical driving scenarios. The results show that the proposed motion planning process has a good real-time performance and adaptability in real traffic scenarios.
1 September 2021 in Solar Energy