Gait phase recognition of multi-mode locomotion based on multi-layer perceptron for the plantar pressure measurement system

The gait phase recognition has some broad application prospects, such as lower limb exoskeleton (LLE). To accurately identify the gait phase in different locomotion modes according to gait patterns and corresponding gait characteristics, we define sets of gait phases and propose a gait phase recognition model using plantar pressure sensing signals. The gait phase recognition algorithm based on the multi-layer perceptron (MLP) is used to study the gait phase recognition in walking and running modes. The experimental results show that the gait phase recognition model can recognize the gait phase in different motion modes based on the plantar pressure sensing information. The gait phase recognition of multi-mode locomotion can provide sufficient control logic reference for the powered exoskeleton robot. Through the data of 1052 gait cycles of 4 participants in the experiment, the accuracy of gait recognition for walking mode is 93.9%, and the accuracy of gait recognition for flying state is 76.5%.