A novel step detection technique for pedestrian dead reckoning based navigation

Abstract

Precise step detection is an essential task for pedestrian dead reckoning based navigation. This study presents a novel step detection method using a Dynamic Weight Integrated Fuzzy C-Means (DWIFCM) algorithm. It utilizes the statistical criteria extracted from the readings of accelerometer sensors and improves the accuracy of step detection. The performance of the proposed step detection technique is evaluated by considering the variations in walking patterns of the pedestrian in different landscapes of the path. The experimental results yield that the proposed method outperforms among the existing benchmark techniques.