关键词: 有路无路相结合的路径规划, 通行成本建模, 时间最优, 高斯混合回归, A^*算法

Abstract:

Path planning plays an important role in emergency rescue and emergency rescue. In these scenarios, vehicles are often able to get faster routes through a combination of off-road and on-road routes. Therefore, a time optimal path planning method based on Gaussian mixture regression and improved A^* algorithm is proposed. First, a time optimal traffic cost model is constructed using the A^* algorithm combined with a vehicle speed coefficient, accounting for various factors affecting vehicle traffic, including road conditions. Second, the Gaussian mixture model is employed to collect trajectory information for the proposed rescue route. Combined with Gaussian mixture regression, this model constrains the search radius of the A^* algorithm, enhancing its search efficiency. Finally, experimental verification is conducted using data from Dengfeng, Henan province. The results show that compared with the four algorithms of 2D A^*, 3D A^*, 2D time optimal A^* and improved time optimal A^* without Gaussian mixture regression constraints, the proposed algorithm reduces the path passage time by 2.02% to 32.31%, decreases code running time by 38.76% to 83.6%, and reduces node traversal by 38.69% to 79.77%. When compared to the recommended path from AutoNavi, the proposed algorithm shortens the path distance by 6.86% to 9.53% and reduces passage time by 8.41% to 17.22%.

Key words: path planning with combination of road and non-road routes, toll cost modeling, optimal timing, Gaussian mixed regression, A^* algorithm