Abstract
Restoration of sensory feedback is a major challenge in lower-limb prosthetics. Traditional prostheses lack real-time somatosensory feedback, leading to gait asymmetry, increased cognitive load, and falls, especially on uneven terrain. This study introduces an adaptive, closed-loop peripheral nerve stimulation (PNS) system that modulates stimulation parameters in real-time based on plantar pressure and terrain slope. We tested the system on three trans-tibial amputees fitted with smart insoles and implanted flat interface nerve electrodes (FINEs) on the sciatic nerve. Dynamic stimulation profiles were adapted using a machine-learning-based terrain identification algorithm. Results showed that closed-loop PNS significantly improved gait symmetry by 18.4%, reduced cognitive load (evidenced by a 22.1% reduction in response time during dual-task walking), and enhanced obstacle negotiation success rates on uneven surfaces compared to the no-stimulation condition. These findings demonstrate that adaptive somatosensory feedback can restore a physiological-like sense of touch and ground awareness, paving the way for safer, more intuitive lower-limb neuroprostheses.
