Improvement of person identification rules based onwalking behavior association between wearable acceleration sensors and intelligent space

This paper proposes a human tracking method with a wearable acceleration sensor and networked laser range scanners in an intelligent space. Human walking activities are measured and associated between laser range scanners and acceleration sensor. In this paper, person identification rules between two types of sensors are improved. Especially, similarity calculation using difference of behaviors and optimal parameters is introduced. Optimal parameters are derived from many tracking results obtained with actual system. Some experimental results with optimal parameters show higher performance of target recognition than cases with the conventional parameters.