Method for Incorporating Path Planning Preferences Based on Layered Cost Map

Path planning is an important research field in mobile unmanned intelligent systems, which is usually divided into global planning and local planning. Among them, common global programming algorithms (such as A^*, Dijkstra, etc.) often only consider collisionless and shortest paths, making it difficult to introduce other semantics and tendencies. In response to the above problems, a method for introducing planning tendencies based on hierarchical cost map is proposed, that is, a planning preference layer is introduced into the traditional cost map environment modeling. The layer extends the specified reference path through the breadth-first search algorithm, making the cost higher the further the path planning is from the reference path, thereby guiding the global planning result towards the reference path. The simulation experiment results show that the path planning preference introduction method based on hierarchical cost map can effectively solve the problems of reference path fitting and path planning oscillation, enabling path planning to better reflect the task intention while meeting the requirements of safety and feasibility. It can provide new ideas for the semantic utilization of intelligent perception results and the deep integration of path planning.