Inertial Navigation Error Correction Based on Map Matching Combining Template Matching and Hidden Markov Model

Inertial Navigation Systems (INS) are widely used in military vehicle positioning. However, integrating acceleration and angular velocity to compute positions leads to cumulative errors. To mitigate the error, we present a map-matching-based INS error correction approach that combines Template Matching (TM) and the Hidden Markov Model (HMM). This approach utilizes the high degree of consistency between the overall features of the INS trajectory and the actual road, treats the trajectory as a template, performs a sliding search on the road network to determine the matching location and correct the lateral and longitudinal errors. In cases where multiple matches exist in the road network, the HMM is applied to optimize the matching process by considering local features. Despite substantial initial errors, our approach demonstrates robust matching and correction capabilities. The experiment of vehicles driving on the road for up to 80km confirms that our approach can reduce the average positioning error from 25.2 meters to 9.95 meters, effectively correcting the errors.