TPRGraphMM: Graph-Based Vehicle Map Matching by Travel Pattern Recognition

This paper presents TPRGraphMM, an innovative map matching model that incorporates travel pattern recognition into Graph-Based Vehicular Map Matching (GraphMM), and utilizes an automated parameter optimization framework to determine the model parameters to enhance the accuracy of map matching. In the field of map matching, adverse natural conditions often lead to data interruptions, resulting in sparse trajectories, which significantly increases the difficulty of accurately mapping trajectory data to the road network. To address this issue, the model proposed in this study offers an effective solution. The study utilizes the dataset collected by Tencent Maps for research purposes. TPRGraphMM has proven to substantially surpass existing methods in accuracy, Longest Common Subsequence (LCS), and precision. Notably, against the GraphMM model, our model achieved accuracy enhancements of 5.6<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>,4.3<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>, and 7.7% at 30-second, 60-second, and 120-second intervals, respectively. Concurrently, LCS improvements were recorded at 3.1%, 2.6%, and 4.4% for the same intervals. These results demonstrate the transformative impact of integrating travel mode recognition into map matching algorithms.