Toward a Computational Model of Constraint-Driven Exploration and Haptic Object Identification

Abstract

A conceptual model of the human haptic system in relation to object identification is presented. The model encompasses major architectural elements including representations of haptically accessible object properties and exploratory procedures (EPs)—dedicated movement patterns that are specialized to extract particular properties. These architectural units are related in processing-specific ways. Properties are associated with exploratory procedures in keeping with the extent to which a given procedure delivers information about a given property. The EPs are associated with one another in keeping with their compatibility, as determined by parameters of motor execution and interactions with the object and the workspace. The resulting architecture is treated as a system of constraints which guide the exploration of an object during the course of identification. The selection of the next step in a sequence of exploration requires that constraints be optimally satisfied. A network approach to constraint satisfaction is implemented and shown to account for a number of previous empirical results concerning the time course of exploration, object classification speed, and incidental learning about object properties. This system has potential applications for robotic haptic exploration.