Posture Activity Categorization and Feature Analysis Using an Artificial Neuromolecular System

Monitoring of posture activities enables accurate differentiation of human behavior. In this paper, an artificial neuromolecular system (ANM), a self-organizing system movtivated from brain information processing, was used to separate human behavior patterns. We also looked into the biometric features of each activity acted by each individual. Five healthy adults were invited to participate in this study. Each individual was asked to perform walking, racewalking, stair ascent/descent and jogging activities ten times. A smart phone was tied up with the left heel of each individual for data collection. Experimental results show that the patterns of heel acceleration uniquely characterize differences for each person's behavior patterns, and the proposed system could be used to analyze and estimate as a classification tool by characteristic differences.