Aiming at the problem of poor adaptability of traditional threshold-based zero-velocity detector in foot-mounted inertial pedestrian navigation system, an interval search zero-velocity detector is designed, which adaptively searches the optimal zero-velocity interval based on the inertial data characteristics of foot motion. Firstly, a gait cycle is considered as the basic unit to search the zero-velocity interval. In a gait cycle, accelerations are mapped to the search space by a nonlinear function to amplify the difference of raw data. Finally, The optimal zero-velocity interval is searched based on the convergence of the acceleration in the search space. Under walking, experiments show that the interval-search method without parameter adjustment has a matching degree of about 97.54% with the SHOE method with the optimal threshold. The positioning accuracy is about 3 m/18 min. Compared with Adaptive-SHOE, interval-search method has similar performance and the matching degree is about 98.51%. In the case of multiple motion mixing, Adaptive-SHOE fails. Interval-search method can still work. Compared with SHOE, the interval-search method has a matching degree of about 89.12% and has better accuracy. Interval-search zero-velocity detector has good engineering application value. © 2020, Editorial Department of Journal of Chinese Inertial Technology. All right reserved.
