Transformer-Based Trajectory Prescription for sEMG-Optimizing Robot-Mediated Rehabilitation

This study introduces a transformer-based model that personalizes robot-mediated rehabilitation by mapping surface electromyography (sEMG) signals to contact force trajectories. Using a transformer encoder-decoder with rotary positional embeddings and flow matching, the model predicts trajectories with less than 8% error on unseen tasks and users, and can be fine-tuned within a single therapy session. Live tests show participants achieve target neuromotor activations with only 15% error, demonstrating the feasibility of activation-targeted trajectory prescription.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

• llms
• rag
• diffusion
• robotics
• fine-tuning