Walking Phase Recognition for People with Lower Limb Disability

This paper presents a total solution on EMG signal-based walking phase recognition for people with lower limb disability. Various environmental factors such as sensed location, walking speed, and ground inclination are taken into consideration in all the phases of signal sensing, feature extraction, feature selection, and classification. Based on analysis on fourteen well-known feature extraction methods in varying environmental situation, this paper proposes a methodology for selecting a good feature set, and then demonstrates effectiveness of the proposed approach with the classification results.