Tactile Sensing
Overview
Tactile sensing studies how to measure contact properties such as force, shape, texture, slip, and material, giving robots information that cameras cannot provide during occlusion or direct interaction. Systems use optical gel sensors, resistive and capacitive arrays, magnetic or fluidic skins, and acoustic or event-based sensors; learning methods range from CNNs and transformers to self-supervised representations and physics-based reconstruction. Researchers evaluate these approaches on force estimation, material recognition, contact localization, slip detection, and manipulation tasks, with growing attention to cross-sensor transfer, tactile simulation, and combining touch with vision, language, or proprioception. A central challenge is making results reliable across sensor designs, contact conditions, and real-world tasks, where calibration, data scarcity, durability, and safety remain important constraints. The field also continues to test when tactile input improves manipulation enough to justify its hardware and integration costs.
Latest papers
139 papers18 papers in the last four weeks, down 14% on the four weeks before. 0.1% of all new papers.