Detangling hair using feedback-driven robotic brushing

摘要

The brushing of hair requires a complex understanding of the interaction between soft hair fibers and the soft brushing device. It is also reliant on having both visual and tactile information. Guided by a recently developed model of soft tangled fiber bundles, we develop a method for optimizing hair brushing by robots which seeks to minimize pain and avoid the build up of jammed entanglements. Using an experimental setup with a custom force measuring sensor and a soft brush end effector, we perform closed-loop experiments on hair brushing of different curliness. This utilizes computer vision to assess the curliness of the hair, after which the hair is brushed using a closed loop controller. To demonstrate this approach hair brushing experiments have been performed on a wide variety of wigs with amount of curl. In addition to hair brushing the insight provided by this model driven approach could be applied to brushing of fibers for textiles, or animal fibers.