This paper presents a low-cost real-time lane-determination system that fuses micro-electromechanical systems inertial sensors (accelerometers and gyroscopes), global navigation satellite system (GNSS), and commercially available road network maps. The system can be used for intelligent transportation systems, telematics applications, and autonomous driving. The system does not depend on visual markings or highly precise GNSS technology, such as DGPS or RTK, and it does not need explicit lane-level resolution maps. High-resolution estimation of the vehicle’s position, velocity, and orientation is implemented by fusing inertial sensors with GNSS in a loosely coupled mode using extended Kalman filter. A curve-to-curve road-level map-matching is implemented using a hidden Markov model followed by a least-square regression step that estimates the vehicle’s lane. The system includes a lane-change detector based on inertial sensors and the filtered vehicle’s state. The system has been realized in real time and tested extensively on real-road data. Experiments showed robust map-matching in challenging road intersections and a 97.14% lane-determination success rate.
A Low-Cost Lane-Determination System Using GNSS/IMU Fusion and HMM-Based Multistage Map Matching
IEEE Transactions on Intelligent Transportation Systems ; 18 , 11 ; 3027-3037
2017-11-01
3378130 byte
Article (Journal)
Electronic Resource
English
A Low-Cost Lane-Determination System Using GNSS/IMU Fusion and HMM-Based Multistage Map Matching
Online Contents | 2017
|GNSS Based Lane Keeping Assist System via Model Predictive Control
SAE Technical Papers | 2019
|GNSS Photo Matching: Positioning using GNSS and Camera in Urban Canyon
British Library Conference Proceedings | 2015
|