Highway Vehicle Route Reconstruction Using Sparse and Noisy Communication Base Station Fingerprints

Abstract

The current application of highway toll system generally uses the Dijkstra algorithm to calculate the shortest path of vehicles from the entrance to the exit to charge. This means that managers have no way of knowing the exact route of vehicles. Also, different routes of highways are often funded and operated by different investors. To address this problem, this article presents a new algorithm to reconstruct trajectories from sparse and noisy fingerprint signals from communication base stations identifications (CBSIDs), with practical application in a high-speed toll collection system in Hubei Province, China. In this solution, we use an inexpensive device that collects signal fingerprint identification numbers from CBSIDs at a low sampling rate. These CBSIDs are then matched with a special CBSID-anchor radiomap, converting the sequence of CBSIDs into a sequence of candidate anchors (toll stations and intersections on highways). Finally, a route mapping algorithm is run to process these candidate anchors and to generate the complete driving route. In the experiment on both simulated and field routes, results show that the proposed algorithm can effectively reconstruct the driving routes of vehicles. The upgraded toll collection system meets the needs of efficient motorway investment, maintenance, and management.