Fast Path Planning of Multienvironmental Factors Comprehensive Constraints in Off-Road Environment Constructed by Aggregation Vector Point Method

The accuracy and efficiency of path planning in off-road environments depend on the construction of off-road environment map information. Previous studies have used the grid method to represent off-road environments, but as the number of grids increases, the path planning time significantly increases. Moreover, environmental factors such as elevation, slope, aspect, and land cover type are all factors that affect the trafficability of autonomous vehicles. Therefore, this article proposes a fast path planning method for multienvironmental factor comprehensive constraints in an off-road environment constructed by the aggregation vector point method. First, the required off-road environment image data are converted into vector points; the trafficability of the point is quantified based on its environmental attribute information, and the impassable points are aggregated. The aggregated vector points are used to construct a vector environment model (VEM). Second, based on the environmental attribute information of the vector points, the slope influence layer, land cover type influence layer, and goal guidance layer are constructed to quantitatively express the comprehensive trafficability of each point. Finally, to verify the effectiveness of the VEM constructed by the proposed aggregation vector method, the Dijkstra algorithm with comprehensive constraints from multiple environmental factors is used to perform path planning in both the VEM and the grid environment model (GEM), which is constructed using the traditional grid method. The results indicate that the VEM constructed using the aggregation vector point method can accurately reflect environmental information while improving path planning efficiency. The path planning time required in grid environments was longer than that in vector environments. In the grid environment of small, medium, and large research areas, the planning path time is about 4.5, 1.9, and 1.26 times that in the vector environment of small, medium, and large research areas, respectively. The path planned by the Dijkstra algorithm after comprehensive constraints is safer and more feasible.