APPLICATION OF CHANGE DETECTION TO DYNAMIC CONTACT SENSING

The forces of contact during manipulation convey substantial informati on about the state of the manipulation. Textures, slip, impacts, grasp ing, and other contact conditions produce force and position signature s that can be used for identifying the state of contact. This article addresses the fundamental problems of interpreting the force signals w ithout any additional context on the state of manipulation. Techniques based on forms of the generalized sequential likelihood ratio test ar e used to segment individual strain signals into statistically equival ent pieces. The results of the segmentation are designed to be used in a higher level procedure that will interpret the results within a man ipulation context. We report on our experimental development of the se gmentation algorithm and on its results for detecting and labeling imp acts, slip, changes in texture, and condition. The sequential likeliho od ratio test is reviewed, and some of its special cases and optimal p roperties are discussed. Finally, we conclude by discussing extensions to the techniques and lessons for sensor design.