Knowledge graph-driven multi-constraint off-road path planning method

Off-road path planning, as an important direction in the field of intelligent navigation, has significant application value in complex scenarios such as emergency rescue, geological exploration, and military deployment. Unlike conventional path planning, which relies on fixed road networks, off-road path planning requires the establishment of a dynamic traversal model to support the planning results. Based on this issue, this study introduces knowledge graphs into the calculation of traversal cost modeling, addressing problems such as low data update efficiency and storage redundancy. First, a multidimensional quantitative analysis model is constructed, including vehicle dynamics parameters, terrain slope characteristics, and surface cover types, to extract the traversal impact factors of various elements through feature engineering. Secondly, based on knowledge graph technology, semantic modeling of multi-source heterogeneous data is achieved, and a dynamic weighted traversal cost map is generated using a rule-based reasoning mechanism. Finally, the performance of path planning is evaluated in a dynamic environment constraint system through a dual-engine validation framework with the improved A* algorithm and Dijkstra algorithm. Experimental results show that, compared to traditional cost map construction methods, our approach improves the construction efficiency in large areas by 30.7%, reduces redundant operations effectively in medium and small areas, and increases efficiency by 75% to 99.4%, which can, to some extent, improve the efficiency of path planning.