The paper reviews different neural network architecture for human gait recognition as a part of the control loop of soft exoskeleton. 6-axis inertial measurement unit (IMU) used for experimental data collection and gait phase prediction. IMU sensors are attached to the hip, ankle and foot of each leg. Three different architecture were chosen for comparison: feed-forward neural network, cascade neural network and recurrent neural network. All neural networks trained on the group of dataset that consist of 100 steps and compared by precision. Dataset was collected during subject walking on a treadmill with a different speed. The experiments show that recurrent neural network more efficient than other one.
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.