L3: Sensing driving conditions for vehicle lane-level localization on highways

When vehicle road-level localization cannot satisfy people's need for convenience and safety driving, lane-level localization becomes a corner stone in Intelligent Transportation System. Existing work on tracking vehicles on lane-level mostly depends on pre-deployed infrastructures and additional hardwares. In this paper, we utilize smartphone sensing of driving conditions for vehicle lane-level localization on highways. We analyze the driving traces collected from real driving environments, finding that each type of lane change has its unique pattern on the vehicle's lateral acceleration. Based on this observation, we propose a Lane-Level Localization (L3) system, which can perform real-time vehicle localization on lane-level only using smartphones when vehicles are driving on highways. Our system first uses embedded sensors in smartphones to capture the patterns of lane change behaviors. Then a Gaussian Distribution is employed to track vehicles on lane-level with tolerance of false detections. Extensive experiments demonstrate that L3 is accurate and robust in real driving environments. The experimental results show that, on average, L3 achieves accuracy of 91.49% on lane change detection and 86.94% on lane-level localization.

I. Introduction

With the development of traffic networks, the demand for more precise vehicle localization becomes very urgent. Compared with road-level, vehicle lane-level localization can solve more traffic problems. Among those problems, driving safety issue is the most severe one. According to statistical analysis, failure to stay in the correct lane causes nearly 17,000 deaths annually, which is about 51 percent of the fatal crashes in USA every year [1]. If driving on improper lanes can be detected and warned, those accidents could be avoided. Another application of lane information is working as the assistance for navigators. It is an ordinary occurrence that a driver who is not familiar with the routes may miss a highway exit when the navigator cannot tell in which lane the driver should get prepared for. If the driver can be alerted for the preparation in time, unnecessary detour could be avoided. Moreover, lane information can also be applied in crowd sensing. If most drivers know which lanes they are driving on and upload the information to a cloud server, it is possible for traffic planers to analyze the conditions of roads on lane-level, e.g., if there is a bad condition on a lane causing a lot of drivers to avoid it, an immediate fixation can be made from traffic institutions. Therefore, it is indispensable to realize vehicle lane-level localization.