No abstract

This article describes a model of motion planning instantiated for grasping. According to the model, one of the most important aspects of motion planning is establishing a constraint hierarchy--a set of prioritized requirements defining the task to be performed. For grasping, constraints include avoiding collisions with to-be-grasped objects and minimizing movement-related effort. These and other constraints are combined with instance retrieval (recall of stored postures) and instance generation (generation of new postures and movements to them) to simulate flexible prehension. Dynamic deadline setting is used to regulate termination of instance generation, and performance of more than one movement at a time with a single effector is used to permit obstacle avoidance. Old and new data are accounted for with the model.

This article describes a theory of the computations underlying the selection of coordinated motion patterns, especially in reaching tasks. The central idea is that when a spatial target is selected as an object to be reached, stored postures are evaluated for the contributions they can make to the task. Weights are assigned to the stored postures, and a single target posture is found by taking a weighted sum of the stored postures. Movement is achieved by reducing the distance between the starting angle and target angle of each joint. The model explains compensation for reduced joint mobility, tool use, practice effects, performance errors, and aspects of movement kinematics. Extensions of the model can account for anticipation and coarticulation effects, movement through via points, and hierarchical control of series of movements.

A goal of research on the cognitive control of movement is to determine how movements are chosen when many movements are possible. We addressed this issue by studying how subjects reached for a bar to be moved as quickly as possible from a home location to a target location. Ss generally grabbed the bar in a way that afforded a comfortable posture at the target location (the end-state comfort effect) and with the thumb toward the end of the bar that would be aligned with the target (the thumb-toward bias). The data suggested that subjects chose handgrips by retrieving instances of previous reaches, not by carrying out computations that treated candidate reaches as new behavioral events.

No abstract

No abstract

Most physical tasks can be performed with an infinite number of movement patterns. How then are particular patterns selected? We propose that the contributions of individual limb segments depend on their own independently assessed fits to task demands. An advantage of this system is that coordination among limb segments can be achieved without explicit control of limb-segment interactions. In addition, the system allows segments that are still functioning to compensate for segments that are disabled. To test the model, we first asked subjects to oscillate the fingertip over varying distances at varying rates, using only the finger, hand, or forearm. Based on their performance, we identified the optimal amplitude and frequency of movement for each limb segment. Then we allowed the subjects to use the finger, hand, and forearm however they wished. We demonstrate that the relative contribution of each limb segment to fingertip displacement is predicted by the similarity of the optimal amplitude and frequency of that segment to the required amplitude and frequency of fingertip displacement. Because our model is similar to models proposed for learning and perception, common computational approaches appear viable for motor control and other more widely studied activities underlying information processing and behavior.

No abstract