Memory Unscented Particle Filter for 6-DOF Tactile Localization

被引:30
作者
Vezzani, Giulia [1 ,2 ]
Pattacini, Ugo [1 ]
Battistelli, Giorgio [3 ]
Chisci, Luigi [3 ]
Natale, Lorenzo [1 ]
机构
[1] Ist Italiano Tecnol, iCub Facil, I-16163 Genoa, Italy
[2] Univ Genoa, I-16145 Genoa, Italy
[3] Univ Florence, Dipartimento Ingn Informaz, I-50121 Florence, Italy
关键词
Bayesian state estimation; particle filtering; tactile localization; KALMAN FILTER; ROBOT; TECHNOLOGIES; OBJECTS; RANGE;
D O I
10.1109/TRO.2017.2707092
中图分类号
TP24 [机器人技术];
学科分类号
080202 ; 1405 ;
摘要
This paper addresses 6-DOF (degree-of-freedom) tactile localization, i. e., the pose estimation of tridimensional objects using tactile measurements. This estimation problem is fundamental for the operation of autonomous robots that are often required to manipulate and grasp objects whose pose is a priori unknown. The nature of tactile measurements, the strict time requirements for real-time operation, and the multimodality of the involved probability distributions pose remarkable challenges and call for advanced nonlinear filtering techniques. Following a Bayesian approach, this paper proposes a novel and effective algorithm, named memory unscented particle filter (MUPF), which solves 6-DOF localization recursively in real time by only exploiting contact point measurements. The MUPF combines a modified particle filter that incorporates a slidingmemory of pastmeasurements to better handle multimodal distributions, along with the unscented Kalman filter that moves the particles toward regions of the search space that are more likely with the measurements. The performance of the proposed MUPF algorithm has been assessed both in simulation and on a real robotic system equipped with tactile sensors (i. e., the iCub humanoid robot). The experiments show that the algorithm provides accurate and reliable localization even with a low number of particles and, hence, is compatible with real-time requirements.
引用
收藏
页码:1139 / 1155
页数:17
相关论文
共 35 条
[31]   A review of tactile sensing technologies with applications in biomedical engineering [J].
Tiwana, Mohsin I. ;
Redmond, Stephen J. ;
Lovell, Nigel H. .
SENSORS AND ACTUATORS A-PHYSICAL, 2012, 179 :17-31
[32]  
Tosi N, 2014, IEEE INT CONF ROBOT, P2270, DOI 10.1109/ICRA.2014.6907173
[33]  
van der Merwe R, 2001, ADV NEUR IN, V13, P584
[34]  
Vezzani G, 2016, IEEE-RAS INT C HUMAN, P256, DOI 10.1109/HUMANOIDS.2016.7803286
[35]   Biomimetic tactile sensor array [J].
Wettels, Nicholas ;
Santos, Veronica J. ;
Johansson, Roland S. ;
Loeb, Gerald E. .
ADVANCED ROBOTICS, 2008, 22 (08) :829-849