Slip Detection on Natural Objects with a Biomimetic Tactile Sensor

Slip detection enables robotic hands to perform complex manipulation tasks by predicting when a held object is about to be dropped. Here we use a support vector machine classifier to detect slip with a biomimetic optical tactile sensor: the TacTip. Previously, this method has been shown to be effective on various artificial stimuli such as flat or curved surfaces. Here, we investigate whether this method generalises to novel, everyday objects. Five different objects are tested which vary in shape, weight, compliance and texture as well as being common objects that one might encounter day-to-day. Success of up to 90% is achieved which demonstrates the classifier’s ability to generalise to a variety of previously unseen, natural objects.

[1]  Hannes P. Saal,et al.  Touch is a team effort: interplay of submodalities in cutaneous sensibility , 2014, Trends in Neurosciences.

[2]  Nathan F. Lepora,et al.  Exploratory Tactile Servoing With Active Touch , 2017, IEEE Robotics and Automation Letters.

[3]  Mark R. Cutkosky,et al.  Sensing skin acceleration for slip and texture perception , 1989, Proceedings, 1989 International Conference on Robotics and Automation.

[4]  Kaspar Althoefer,et al.  Tactile sensing for dexterous in-hand manipulation in robotics-A review , 2011 .

[5]  Jonathan Rossiter,et al.  The TacTip Family: Soft Optical Tactile Sensors with 3D-Printed Biomimetic Morphologies , 2018, Soft robotics.

[6]  Jonathan Rossiter,et al.  Development of a tactile sensor based on biologically inspired edge encoding , 2009, 2009 International Conference on Advanced Robotics.

[7]  Nathan F. Lepora,et al.  Slip Detection With a Biomimetic Tactile Sensor , 2018, IEEE Robotics and Automation Letters.

[8]  Jan Peters,et al.  Stabilizing novel objects by learning to predict tactile slip , 2015, 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[9]  J. Randall Flanagan,et al.  Coding and use of tactile signals from the fingertips in object manipulation tasks , 2009, Nature Reviews Neuroscience.