Continuous Motion Recognition for Natural Pedestrian Dead Reckoning Using Smartphone Sensors

Jiuchao Qian, Ling Pei, Rendong Ying, Xin Chen, Danping Zou, Peilin Liu, Wenxian Yu

Abstract:	Among various indoor navigation technologies, Pedestrian Dead Reckoning (PDR) is an algorithm based on Magnetic Angular Rate and Gravity (MARG) sensor arrays including accelerometers, gyroscopes and magnetometers which are gradually integrated into smartphones along with the continuous miniaturization of micro electromechanical system (MEMS). Usually, PDR is adopted to provide means of reducing the inertial error accumulation arises from sensor noise and drift by taking advantage of the sequential characteristics of pedestrian motion. And the algorithm is a relative navigation technique, which determines the relative location of a pedestrian by using step detection, stride length estimation, and heading determination. Typically, the accelerometer measurements are utilized to carry out step detection and stride length estimation, and heading determination is simultaneously completed by fusing the information from gyroscopes and magnetometers. In this paper, a continuous motion recognition algorithm for natural pedestrian dead reckoning has been proposed. To address the challenge of PDR systematic errors arises from the natural walk motion, individual motion features are researched and complex motions are segmented into a series of basic motion components, thereby continuous pedestrian motion recognition is accomplished. First of all, this paper focuses on modeling various basic motion states during the natural walk process including four broad categories: static state, walking state, turning state and stairs state. Then, in terms of observation sequences of basic motion states, we perform sequence labeling and segmentation of complex motions simultaneously from a continuous stream of basic motions using Conditional Random Fields (CRFs) algorithm. Finally, adaptive stride length estimation and heading determination are implemented to improve the performance of the PDR system. Experimental results show that the proposed algorithm achieves accuracy comparable to previous results gained under ideal experimental conditions.
Published in:	Proceedings of the 27th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2014) September 8 - 12, 2014 Tampa Convention Center Tampa, Florida
Pages:	1796 - 1801
Cite this article:	Qian, Jiuchao, Pei, Ling, Ying, Rendong, Chen, Xin, Zou, Danping, Liu, Peilin, Yu, Wenxian, "Continuous Motion Recognition for Natural Pedestrian Dead Reckoning Using Smartphone Sensors," Proceedings of the 27th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2014), Tampa, Florida, September 2014, pp. 1796-1801.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In