A Historical-Trajectories-Based Map Matching Algorithm for Container Positioning and Tracking

• Published in: Sensors (Basel, Switzerland) Vol. 22; no. 8; p. 3057
• Main Authors: Li, Wenfeng, Zhang, Wenwen, Gao, Cong
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 15.04.2022; MDPI
• Subjects: Accuracy; Algorithms; container positioning and tracking; historical trajectories; Internet of Things; Low earth orbit satellites; map matching; path reconstruction; Probability; Probability distribution; path reconstruction; historical trajectories; map matching; container positioning and tracking
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s22083057

Positioning and tracking of containers is becoming an urgent demand of container transportation. Map matching algorithms have been widely applied to correct positioning errors. Because container trajectories have the characteristics of low sampling rate and missing GPS points, existing map matching algorithms based on the shortest path principle are not applicable for container positioning and tracking. To solve this problem, a historical-trajectories-based map matching algorithm (HTMM) is proposed. HTMM mines the travel time and the frequency in historical trajectories to help find the local path between two adjacent candidate points. HTMM first presents a path reconstruction method to calculate the travel time of historical trajectories on each road segment. A historical path index library based on a path tree is then developed to efficiently index historical paths. In addition, a location query and tracking method is introduced to determine the location of containers at given time. The performance of HTMM is validated on a real freight trajectory dataset. The experimental results show that HTMM has more than 3% and 5% improvement over the ST-Matching algorithm and HMM-based algorithm, respectively, at 60–300 s sampling intervals. The positioning error is reduced by half at a 60 s sampling interval.