Using a fascinating combination of ridges like fingerprints for fine-grained texture sensing, plus a fluid casing for its vibrotactile sensors, this is a fascinating approach from USC to the haptic perception of textures. Apparently 95% accurate, judged against a database of textiles, this is better than human assessment. As with all forms of sensory perception, the biggest difficulty lies between the human’s ears. Published in Frontiers in Neurorobotics.
Robots get a feel for the world (USC)
Sensory/Motor 
cool post Mark, this level of sensitivity in a machine is remarkable and could/will allow robotics intimate access to humans – what next?