A Hybrid Intelligent Multisensor Positioning Methodology for Reliable Vehicle Navigation

With the rapid development of intelligent transportation systems worldwide, it becomes more important to realize accurate and reliable vehicle positioning in various environments whether GPS is available or not. This paper proposes a hybrid intelligent multisensor positioning methodology fusing the information from low-cost sensors including GPS, MEMS-based strapdown inertial navigation system (SINS) and electronic compass, and velocity constraint, which can achieve a significant performance improvement over the integration scheme only including GPS and MEMS-based SINS. First, the filter model of SINS aided by multiple sensors is presented in detail and then an improved Kalman filter with sequential measurement-update processing is developed to realize the filtering fusion. Further, a least square support vector machine- (LS SVM-) based intelligent module is designed and augmented with the improved KF to constitute the hybrid positioning system. In case of GPS outages, the LS SVM-based intelligent module trained recently is used to predict the position error to achieve more accurate positioning performance. Finally, the proposed hybrid positioning method is evaluated and compared with traditional methods through real field test data. The experimental results validate the feasibility and effectiveness of the proposed method.