An enhanced weight-based real-time map matching algorithm for complex urban networks

A map-matching algorithm is used to map the inaccurate raw coordinate data to the digital road network. It is an indispensable part of Location Based Service applications and Intelligent Transportation Systems, such as navigation systems. Accuracy and performance (running speed) are usually traded off in traditional algorithms. An enhanced weight-based real-time map matching algorithm only employing GPS data is proposed to guarantee both. The algorithm has two steps: initialization and tracking match, each step is mainly composed of three parts. Firstly, segments near the GPS point are selected as candidate segments. Secondly, four criteria (distance, heading difference, direction difference and segment connectivity) are used to identify the best segment among candidates. Considering the reliability of each criterion, four dynamic weight coefficients are introduced. Finally, before assigning a candidate segment to each GPS point, a confidence level is calculated and considered based on the density and complexity of roads around the point. We evaluate the algorithm with field data collected from the city of Chongqing, China. The results demonstrate that it can identify correct segment from complicated and dense urban road networks, with an average matching accuracy of 97.31% and a latency of 3.20ms per location estimate. (C) 2019 Published by Elsevier B.V.