TY - JOUR
T1 - Bayesian Tactile Exploration for Compliant Docking with Uncertain Shapes
AU - Hauser, Kris
N1 - Dr. Hauser received the Stanford Graduate Fellowship, Siebel Scholar Fellowship, Best Paper Award at IEEE Humanoids in 2015, and an NSF CAREER Award.
Manuscript received November 5, 2018; accepted May 18, 2019. Date of publication July 3, 2019; date of current version October 1, 2019. This paper was recommended for publication by Associate Editor F. Kanehiro and Editor E. Yoshida upon evaluation of the reviewers’ comments. This work was supported by the National Science Foundation National Robotics Initiative Grant 1527826.
PY - 2019/10
Y1 - 2019/10
N2 - This paper presents a Bayesian approach for active tactile exploration of a planar shape in the presence of both localization and shape uncertainty. The goal is to dock the robot's end-effector against the shape-reaching a point of contact that resists a desired load-with as few probing actions as possible. The proposed method repeatedly performs inference, planning, and execution steps. Given a prior probability distribution over object shape and sensor readings from previously executed motions, the posterior distribution is inferred using a novel and efficient Hamiltonian Monte Carlo method. The optimal docking site is chosen to maximize docking probability, using a closed-form probabilistic simulation that accepts rigid and compliant motion models under Coulomb friction. Numerical experiments demonstrate that this method requires fewer exploration actions to dock than heuristics and information-gain strategies.
AB - This paper presents a Bayesian approach for active tactile exploration of a planar shape in the presence of both localization and shape uncertainty. The goal is to dock the robot's end-effector against the shape-reaching a point of contact that resists a desired load-with as few probing actions as possible. The proposed method repeatedly performs inference, planning, and execution steps. Given a prior probability distribution over object shape and sensor readings from previously executed motions, the posterior distribution is inferred using a novel and efficient Hamiltonian Monte Carlo method. The optimal docking site is chosen to maximize docking probability, using a closed-form probabilistic simulation that accepts rigid and compliant motion models under Coulomb friction. Numerical experiments demonstrate that this method requires fewer exploration actions to dock than heuristics and information-gain strategies.
KW - Climbing robots
KW - force and tactile sensing
KW - motion and path planning
KW - probability and statistical methods
UR - http://www.scopus.com/inward/record.url?scp=85077760587&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85077760587&partnerID=8YFLogxK
U2 - 10.1109/TRO.2019.2921144
DO - 10.1109/TRO.2019.2921144
M3 - Article
AN - SCOPUS:85077760587
SN - 1552-3098
VL - 35
SP - 1084
EP - 1096
JO - IEEE Transactions on Robotics
JF - IEEE Transactions on Robotics
IS - 5
M1 - 8755306
ER -