Representation framework of perceived object softness characteristics for active robotic hand exploration


Representation framework of perceived object softness characteristics for active robotic hand exploration


Ricardo Martins [ website | social ]

Diego R. Faria [ website | social ]

Jorge Dias [ website | social ]


During the last years the principles and demands guiding the design and implementation of robotic platforms are changing. Nowadays, robotic platforms are tendency equipped with a conjugation of multi-modal artficial perception systems (stereo vision, tactile sensing) and complex actuation systems (multi-articulated robotic hands, arms and legs). This artificial perception systems are required by robotic systems to navigate and interact with the environment and persons. This work is focused in the artificial perception systems related with the robotic manipulation strategies used to dexterously interact with deformable objects in the environment.In this context, the robotic dexterous manipulation of objects require that the framework used to represent the characteristics of the deformable manipulated objects should besuitable to receive inputs from multiple exploratory elements (multi-fingered robotic hands) and to progressively update that representation status as long as the exploration progresses on time. The framework should be also designed to incorporate the uncertainty and errors associated to the sensing process in this type of dynamic environments, and to deal with novelty by characterizing objects of new softness characteristics to the system based on the previous knowledge and interactions with a restricted set of reference materials, that constitute the haptic memory of the system.In order to provide to the robotic hands the capability to differentiate deformable objects with distinct softness characteristics and to dexterously manipulate them, this work analyses the principles and strategies used by humans to sucessfully perform such type of tasks, using predominately haptic information. During the object exploration, the perception and discrimination capability of softness characteristics depend on both cutaneous and kinesthetic information by executing press and release movements active haptic perception. This has been demonstrated by psychophysical experiments performed by Srinivasan and Lamotte.This work intends to contribute to the development of autonomous dexterous robotic hands platforms by proposing a probabilistic volumetric framework to represent and discriminate the perceived softness characteristics extracted during the exploration of deformable objects. The main contributions of this work are the model used to represent the perceived interaction signature of the active exploratory elements and the reference materials with different softness characteristics; the methods used to update the perceived softness volumetric representation of the explored object (Integration and fusion of the cutaneous and kinesthetic data, integration of the local context information).


7th ACM/IEEE International Conference on Human Robot Interaction (HRI2012) - Workshop on Advances in Tactile Sensing and Touch based Human-Robot Interaction

Date Published:


Alternative full-text PDF:

download full-text PDF via University of Coimbra


robotics, artificial perception, Bayesian modelling, haptic exploration, probabilistic grid maps



Representation framework of perceived object softness characteristics for active robotic hand exploration