A novel heading estimation algorithm for pedestrian using a smartphone without attitude constraints

摘要

Pedestrian Dead Reckoning is an algorithm which uses sensors such as accelerometer, gyroscope and magnetometer to provide real-time location services. Since these sensors are integrated into smartphones, smartphone based PDR system becomes an efficient alternative of indoor positioning method. Heading estimation is a crucial part of PDR system, because step length of a pedestrian can be well modelled and a slight deviation may result in huge cumulative positioning error. Several algorithms have been proposed for heading estimation, but they need the smartphone to keep relatively stationary to the pedestrian, which is inconvenient and difficult to popularize. To remedy these disadvantages, we provide a novel method of estimating moving direction of the pedestrian using data from accelerometer. It allows different smartphone holding postures and relative motion between the pedestrian and his phone. After applying this method, it's unnecessary for pedestrian to maintain the same attitude angles of the smartphone for positioning. The experiment result shows that this method performs robust and accurate results under different smartphone holding postures. The PDR assisted with this method makes a good performance of positioning under different smartphone holding postures.