In a new study announced on 19 Nov 2025, Georgia Tech researchers demonstrated a system that uses machine-learning models to dynamically adapt exoskeleton behaviour to individual users in minutes rather than hours. The system utilises real-time motion capture, reinforcement learning and embedded sensors to minimise human-robot mismatch.
According to lead researcher Dr. R Sundar (not the entire name), this approach reduces the barrier to deploying wearable robotic assistants in industrial and healthcare settings. Potential applications include mobility aids, worker-assist devices and rehabilitation systems. The breakthrough suggests that the practical adoption of exoskeletons is accelerating beyond lab prototypes.
Real-World Helper Exoskeletons Take a Step Forward at Georgia Tech
