Deceleration during braking could be controlled by (a) using the time derivative of the relative rate of optical expansion, relative to a -0.5 margin value of tau-dot (D.N. Lee, 1976) or (b) computing the required deceleration from spatial variables (i.e., perceived distance, velocity, or object size). Participants viewed closed-loop displays of approach to an object and regulated their deceleration with a brake. The object appeared on a checkerboard ground surface (providing velocity, distance, and size information) or with no background (providing only optical expansion). Mean tau-dot during braking was -0.51, and estimates of the critical value of tau-dot based on brake adjustments were -0.44 and -0.52, close to the expected value. There were no overall effects of the ground surface or object size. The results are consistent with a tau-dot strategy, where the direction and magnitude of brake adjustments are regulated using tau-dot.

Three experiments examined the functional specificity of visually controlled posture during locomotion by presenting large-screen displays to participants walking on a treadmill. Displays simulated locomotion down a stationary hallway, a hallway that traveled with the observer, or a frontal wall that traveled with the observer. A superimposed oscillation specified postural sway in 6 possible directions. With the wall, sway amplitude was isotropic and directionally specific in all conditions. However, with the hallways, sway was anisotropic (lateral > anterior-posterior [AP]), and diagonal responses were flattened into the lateral plane. When the treadmill was turned 90 ° to the hallway, both the anisotropy and flattening were reversed (AP > lateral), indicating that they are determined by the visual structure of the scene. The results can be explained by postural control laws based on both optical expansion and motion parallax, yielding biases in planar environments that truncate parallax.It is often supposed that the visual control of locomotion is based on optic flow patterns produced at the eye of a moving observer (Gibson, 1950;Lee, 1974;Warren, Morris, & Kalish, 1988). However, there is little direct evidence that human locomotion is actually regulated by such information. Here we report the first in a series of studies that examine how optic flow is used to control posture and gait. In this article, we examine postural responses to optical oscillations during walking. An unexpected pattern of biases in compensatory sway provides a window into the visual control laws for posture. Laws of ControlThe control of locomotion exemplifies the general problem of adaptive visual control. A standard view in psychology, artificial intelligence, and neuroscience has been that various types of information are used to construct a generalpurpose three-dimensional (3D) representation of the environment, on the basis of which actions are planned. Although this model-based approach provides generality, its success has been limited by the computational problems of constructing a sufficiently accurate 3D model from visual data and using it to regulate a many-degrees-of-freedom system in real time.Alternatively, a task-specific approach capitalizes on the regularities of a particular task, yielding special-purpose control relations between informational variables and the free parameters of an action system that is organized for the

No abstract

No abstract

No abstract