Title: Research on Dual-MIMU Trajectory Tracking Based on Support Vector Machine Constraint
Author(s): Qiuying Wang, Ming Cheng, Xufei Cui, Zheng Guo, Jia Li
Published in: Proceedings of IEEE/ION PLANS 2018
April 23 - 26, 2018
Hyatt Regency Hotel
Monterey, CA
Pages: 557 - 564
Cite this article: Wang, Qiuying, Cheng, Ming, Cui, Xufei, Guo, Zheng, Li, Jia, "Research on Dual-MIMU Trajectory Tracking Based on Support Vector Machine Constraint," Proceedings of IEEE/ION PLANS 2018, Monterey, CA, April 2018, pp. 557-564.
Full Paper: ION Members/Non-Members: 1 Download Credit
Sign In
Abstract: Using foot-mounted inertial sensor to track the trajectory is one of the main methods for indoor autonomous pedestrian navigation. However, due to its algorithm, inertial error accumulates over time. To solve this problem, the zero velocity update (ZUPT) takes the velocity error as the observation to amend other information of the carrier. Because zero velocity update can only observe the information of velocity and two horizontal angles, we cannot obtain the position information. In this paper, a dual-MIMU trajectory tracking based on support vector machine(SVM) constraint method is proposed. According to the motion of the human body, in the process of walking, the horizontal angular velocity is the greatest when the toes are off the ground. The theory of s SVM is used to collect the angular velocity of foot movement to classify the movements during the traveling process, Increase the number of observations, after observability analysis to improve positioning accuracy. According to the constraint of the largest step in walking, the inequality equation is constructed and a Kalman filter algorithm is designed. The average position error is 1.3% after 5min’s walking. It verifies that this system has higher positioning accuracy.