Body Motion Based Personal Dead-reckoning System And Algorithm Study

Navigation Positioning and navigation technology play an important role in modern society. Currently the most popular navigation technology is GPS, which features high precision and wide coverage. However, GPS signal is easy to be blocked. So a new positioning technology is needed in the GPS-denied environments, i.e., indoor or underground environments, in the streets surrounded by skyscrapers, in dense forests or complex terrain valleys, etc. Motion sensing based personal positioning technology has advantages of little environmental constraints, flexibility in use, and good positioning robustness. It has important theoretical and practical significance for better responding to personal navigation and positioning tasks in complex environment. For personal positioning applications, especially for those first-responders such as firefighters, emergency rescuers etc, this technology is of extremely high value.This paper firstly designs a personal dead-reckoning system and platform. This system is ARM9-based embedded system platform. Sensors signal acquisition circuits, the GPS input modules interface and results output module design and development are done on the ARM9-based platform. This system has been tested, has good performance, can be used for studying personal dead-reckoning algorithm.Step length and heading are the two important factors in the personal dead-reckoning. For the estimate of the motion step, this paper presents a novel step length estimation model in line with the characteristic of human motion. With this model the step length can be obtained in real time from measuring the step frequency. For step detection, this paper proposes a new step count method by using single-axis gyroscope, based on the traditional step count method by using accelerometer. For the estimate of heading, this paper presents a heading estimate method by fusing gyroscope and electronic compass. When step length and heading are in hand, dead-reckoning can be done step by step. This dead reckoning algorithm has good positioning accuracy and stability. In order to correct the accumulated error of dead reckoning, the paper fuses the GPS and dead-reckoning, when GPS signal is available, using GPS positioning as the position result. Experiments show that the fusion method has a better performance in the environment where GPS signal is good.Finally, in order to further improve the accuracy of the dead reckoning, as well as the accuracy and applicability of the fusing GPS and dead-reckoning navigation, the paper presents fusion algorithm based on Kalman filtering. For more accurate heading, an orientation-stablizing method, where a piece-wise averaging filter was applied upon the Kalman filtering results from gyro and compass data, was presented to further filter out the orientation varies caused by the swing of human body. In order to further reduce the accumulated error of pure dead reckoning, the paper presents a combination positioning algorithm by fusing GPS and dead-reckoning based on extended Kalman filter. Experimental results show that the proposed dead reckoning method has satisfactory performance and the position estimation algorithm fusing intermittent GPS further improve the positioning accuracy.