Automated gait discrimination using Hidden Markov Model

An automated gait pattern discrimination method based on Hidden Markov Model (HMM) was proposed in this paper. According to human gait process, the acceleration signals of human lower limb were divided into different segments and the gait features was extracted by wavelet transform. Then, each state of gait was matched with HMM state and the HMM of each gait mode was trained according to the sample. When the parameters of model are stable, we use the trained HMM to recognize acceleration feature and get the gait pattern ultimately. The experimental results show that HMM has a unique advantage in classification and recognition of timing varying signal.