Adaptive Step Length Method of Dead Reckoning

Aimed at the demand of positioning in complicated environments such as indoor and built-up areas, the method of autonomous navigation positioning and orientation based on an adaptive step length model was proposed in this paper. Using an accelerometer placed on the human waist, stride was recognized and the relationship between the step length, stride factor, and walking speed was established. The stride factor was calibrated by least squares, which adaptively adjusted the calculation of the step length in situations of different walking speeds and manners, thus enhancing the accuracy of the step length. Additionally, information of the magnetometers, gyroscopes, and other sensors were fused to estimate heading with an extended Kalman filter. Personal dead reckoning was achieved based on the results of the moving distance and heading. The experimental results showed that positioning errors of the method were less than 3.5% of the travel distance, and it was able to meet the positioning requirements of pedestrians.