Research on human gait recognition based on fractal theory

Abstract

In order to control the exoskeleton robot effectively, movement capture of the lower limbs and gait recognition is required. Joint angle change contains huge amount of motion information during the movement of human lower limbs, so that optical fiber can be used as an effective method to measure the joint angle of lower limb to achieve gait recognition. In this paper, fractal theory is used for feature extraction based on measured limb angle data. Then, Support Vector Machine (SVM) which has brilliant classification performance for small sample size problems is used for pattern recognition. At last, six motions of human lower limb are identified by the system which are walk, run, squat, stand, uphill and downhill for specific. Experimental results show that the proposed method has distinguished recognition performance.