A navigation path and high-definition map matching scheme based on improved hidden Markov model
Haiyan Liu, Kunfeng Wang, Yadong Wang · 2022 34th Chinese Control and Decision Conference (CCDC) · 2022
In order to realize the autopilot function with the navigation path, it is necessary to match the navigation path output from the navigation module with the road (road center line) of the high-definition map. Generally speaking, the absolute position accuracy of the high-definition map is at the centimeter level, and the absolute position accuracy of the navigation map is at the meter level, while the navigation path is generated on the navigation map, so the navigation path and the road of the high-definition map cannot be matched directly by the position information. In this paper, we propose a scheme for matching navigation paths and high-definition maps based on an improved Hidden Markov Model. Combining the global longitude and latitude positions of navigation path points and the attribute information on the path for initial matching, the navigation path is matched with the center line of the road of the high- definition map. And during the vehicle navigation, the initial matching results are checked using the navigation events within two kilometers ahead sent by the navigation, and the matching errors and missed matching roads are corrected to further improve the matching effect. Finally, this paper proves the effectiveness of this scheme by analyzing the typical evaluation results of initial matching.