Psychophysics of the hand

PhD vacancy
Title of the project: Psychophysics of the hand
Funding:  EU (PhD)

Postdoc vacancy
Title of the project: Psychophysics of the hand
Funding:  EU (Postdoc)

 

Description:

Start of project: Early 2010

In this European project, entitled The Hand Embodied (THE), groups from Italy, Germany, Sweden, France, Greece and The Netherlands (Utrecht) will work together. By learning from human data and hypotheses-driven simulations how to devise improved system architectures for the “hand” as a cognitive organ, the ultimate goal of the project is to design and control new and improved robot hands, haptic interfaces, and hand prostheses.

The Human Perception group in Utrecht, headed by prof. Astrid Kappers, will be responsible for a number of subprojects. These subprojects will be divided over the postdoc(s) and PhD-student(s) that will be hired for this project. These subprojects include the following:

Feature saliency
One of the questions that needs to be answered is which features are salient for perception. For example, shape is an important cue for recognizing an object by touch. Several features, such as edges, curvature, surface area, and aspect ratio, are associated with 3-D shape. To investigate the saliency of 3-D shape features, a haptic search task, an adaptation of a task widely used in visual perception [Wolfe 98], offers a whole new perspective [Plaisier 09]. Participants have to grasp stimuli and respond as quickly as possible whether or not an object with a certain characteristic is present in the stimulus set. Results show that this type of haptic search task can be performed very efficiently (25 msec/item). So far, only geometric features have been studied, showing that edges and vertices are the most salient features. In this workpackage, the saliency of all kinds of object and material properties, such as geometrical shape, temperature, heat conductance, roughness, compliance, etc. will be investigated.

Perceptual integration: haptic multi-cue
This subproject focuses on how shape is perceived through grasping. Integration of the information from multiple contact points is necessary for obtaining a shape. When you grasp a cylinder, you are quite accurate in determining whether or not its cross-section is circular [Van der Horst 08]. In fact, thresholds in terms of aspect ratios are much smaller for cylinders than for rectangular blocks, which suggests that you are apparently quite sensitive to changing curvature over the object and not to aspect ratio per se. This holds true for both dynamic and “static” grasping of the stimulus. On the other hand, studies on the discrimination of curvature show that thresholds are mainly determined by first order information, i.e. slope differences over the stimulus [Pont 99, Wijntjes in press]. Here, this research will be extended to shapes that vary in three dimensions, for example, spheres and ellipsoids, cubes and rectangular blocks.
Another approach will be to investigate through measuring and modelling, how accurate the local information needs to be in order to be able to make a reliable decision given certain classes of stimuli. Relevant questions are: How accurate do humans need to know the positions of their fingers relative to each other? Does edge information (as in the case of rectangular blocks) add so much noise to the stimulation that accuracy deteriorates? Are subjects able to accurately estimate the cross-section by just grasping the object? etc.

Exploratory procedures
A sequel to these studies will be to add movements. Humans are very skilled in using their hands for all kinds of explorations and actions. They are able to grasp and explore objects and decide whether they are warm or cold, smooth or rough, heavy or light, big or small, slippery, etc. Lederman and Klatzky [Lederman 87] studied these movements and they were able to classify them in what they termed a set of exploratory procedures (EPs). In specific tasks (for example, determining the shape of an object), specific EPs will lead to optimal performance. Although this work is highly cited, so far quantitative analyses of the hand movements have not been done. Recently, it has been found that although the hand has a complex structure, in many tasks such as grasping or typing, the movements of the joints are correlated [Thakur 08]. By means of principal component analysis these researchers could identify a number of synergies that were remarkably similar over subjects and objects. An interesting question that will be addressed in this project, is whether EPs can be linked to synergies. One obvious technique that will be used is principal component analysis of hand movements, but other techniques will be considered as well..

 

If you are interested in joining the Human Perception group to work on this ambitious, international project, please contact prof dr. Astrid Kappers, preferably by e-mail.

---