SmartStim: A Recurrent Neural Network Assisted Adaptive Functional Electrical Stimulation for Walking

Abstract According to the Neuro Patience report of Neurological Alliance, 1 in 6 people UK has a neurological condition. With growth technology, rehabilitation for problems is one fastgrowing fields. Functional Electrical Stimulation (FES) those neuro-rehabilitation methods that uses electrical nerve stimulation restore functional muscle movements are lost due such as stroke and multiple sclerosis. This neuroprosthetic device frequently used assist walking by treating condition called Drop Foot, result paralysis pretibial muscles. study proposes two-channel FES SmartStim, which ability modulate its levels according various obstacles stairs ramps. system employs sensor-based module with Recurrent Neural Network classify these different scenarios. The built Inertial Measurement sensors embedded pair shoes, data from predict user walking. These predictions then Fuzzy Logic Controller control regulate current two channels SmartStim system. In system, channel will help aid drop foot, while other be stimulate another group access ramps user. this been trained tested using k-fold cross-validation. evaluation model shows it can sensor at 97 percent accuracy. Currently, further testing being performed assess workings fuzzy logic controller combination healthy individuals. It expected may users accessing scenarios more efficiently.