Robot Assisted Dressing

Robot-assisted dressing aims to develop robotic systems capable of helping individuals with disabilities or limited mobility dress themselves, improving their independence and quality of life. Current research focuses on improving the accuracy and safety of robot manipulation of clothing, employing techniques like model predictive control, diffusion models for cloth state estimation, and reinforcement learning to train policies that handle diverse body poses and garment types. These advancements leverage multi-modal sensing (vision and force) and often incorporate physics-based simulations to improve generalization and safety, ultimately aiming to create robust and reliable assistive robots for this challenging task.