Haptic Object Identification for Advanced
Manipulation Skills
Volker Gabler
1(B) , Korbinian Maier
1 , Satoshi Endo
2 ,
and Dirk Wollherr
1
1 Chair of Automatic Control Engineering, Munich, Germany
{v.gabler,korbinian.maier,dw}@tum.de
2 Chair of Information Oriented Control, Department of Electrical and Computer
Engineering, Technical University of Munich, Munich, Germany
s.endo@tum.de
Abstract. In order to identify the characteristics of unknown objects,
humans - in contrast to robotic systems - are experts in exploiting their
sensory and motoric abilities to refine visual information via haptic perception. While recent research has focused on either estimating the geometry or material properties, this work strives to combine these aspects by
outlining a probabilistic framework that efficiently refines initial knowledge from visual sensors by generating a belief state over the object shape
while simultaneously learn material parameters. Specifically, we present
a grid-based and a shape-based exploration strategy, that both apply the
concepts of Bayesian-Filter theory in order to decrease the uncertainty.
Furthermore, the presented framework is able to learn about the geometry as well as to distinguish areas of different material types by applying
unsupervised machine learning methods. The experimental results from
a virtual exploration task highlight the potential of the presented methods towards enabling robots to autonomously explore unknown objects,
yielding information about shape and structure of the underlying object
and thus, opening doors to robotic applications where environmental
knowledge is limited.
Keywords: Haptic identification · Object classification · Autonomous
agents
In order to allow robots to manipulate arbitrary objects, such as electronic
waste components for automated disassembly, object identification and knowledge acquisition is crucial. While a rough estimation of the shape of an object
can be obtained from visual data, the exact material decomposition remains in
general unknown. Nonetheless, these material properties have a distinct effect on
the selection of the subsequent manipulation tasks, e.g. the choice of materialdependent cutting tools. In order to allow robots to autonomously identify these
object characteristics, one approach is to mimic human behavior in applying haptic data acquisition methods, i.e. actively interacting with the unknown object.
c
Springer Nature Switzerland AG 2020
V. Vouloutsi et al. (Eds.): Living Machines 2020, LNAI 12413, pp. 128–140, 2020.
https://doi.org/10.1007/978-3-030-64313-3_14
Manipulation Skills
Volker Gabler
1(B) , Korbinian Maier
1 , Satoshi Endo
2 ,
and Dirk Wollherr
1
1 Chair of Automatic Control Engineering, Munich, Germany
{v.gabler,korbinian.maier,dw}@tum.de
2 Chair of Information Oriented Control, Department of Electrical and Computer
Engineering, Technical University of Munich, Munich, Germany
s.endo@tum.de
Abstract. In order to identify the characteristics of unknown objects,
humans - in contrast to robotic systems - are experts in exploiting their
sensory and motoric abilities to refine visual information via haptic perception. While recent research has focused on either estimating the geometry or material properties, this work strives to combine these aspects by
outlining a probabilistic framework that efficiently refines initial knowledge from visual sensors by generating a belief state over the object shape
while simultaneously learn material parameters. Specifically, we present
a grid-based and a shape-based exploration strategy, that both apply the
concepts of Bayesian-Filter theory in order to decrease the uncertainty.
Furthermore, the presented framework is able to learn about the geometry as well as to distinguish areas of different material types by applying
unsupervised machine learning methods. The experimental results from
a virtual exploration task highlight the potential of the presented methods towards enabling robots to autonomously explore unknown objects,
yielding information about shape and structure of the underlying object
and thus, opening doors to robotic applications where environmental
knowledge is limited.
Keywords: Haptic identification · Object classification · Autonomous
agents
In order to allow robots to manipulate arbitrary objects, such as electronic
waste components for automated disassembly, object identification and knowledge acquisition is crucial. While a rough estimation of the shape of an object
can be obtained from visual data, the exact material decomposition remains in
general unknown. Nonetheless, these material properties have a distinct effect on
the selection of the subsequent manipulation tasks, e.g. the choice of materialdependent cutting tools. In order to allow robots to autonomously identify these
object characteristics, one approach is to mimic human behavior in applying haptic data acquisition methods, i.e. actively interacting with the unknown object.
c
Springer Nature Switzerland AG 2020
V. Vouloutsi et al. (Eds.): Living Machines 2020, LNAI 12413, pp. 128–140, 2020.
https://doi.org/10.1007/978-3-030-64313-3_14