Walking Status Detection for Pedestrian Navigation

Abstract

Pedestrian Navigation System (PNS) is useful for emergency responders, security personnel, and wide range augmented reality applications. Since the majority of these pedestrian navigation applications require walking from outdoor to an indoor destination point, the Inertial navigation System (INS) which is self-contained and immune to jamming/interference are preferred to be used. Due to the sensor properties, a navigation solution of a stand-alone inertial system will drift rapidly, so it relies on external corrections to maintain the system stability and reliability, such as Global Positioning System (GPS). In addition to seeking for external compensation, taking use of constraints deduced from the particular properties of pedestrian application is also well presented. Compared with using external augmentation, these constraints are advantaged of keeping the system independence. In this paper, the walking status detection based on inertial sensor’s properties is investigated. Two constraints are considered to reduce the divergence of the navigation solution caused by the signal errors of accelerometer and gyroscope. One is position update by step detection and the other is the heading angle constraint based on walking status detection. Experiments results show that different walking status: static, walking forward, or turning, can be well distinguished by the signal properties during walking.