Over the past four years we have been studying problems of control and coordination with funds from AFOSR. Work over the first three years has been reviewed in the final report for the original AFOSR award (Flach, 1996). This report reviews work for the last year of the ASSERT grant. This work has focused on perception and control of low altitude flight. Experimental work from two synthetic task environments is described in this report. The first environment involved descent to low altitude. The key independent variables were speed of forward motion and optical texture (dot, grid, splay, depression). Results showed an interaction between texture and speed. For textures that contained depression information, the rate of approach to asymptote decreased with increasing forward speed. This was not true for splay texture. These results are consistent with previous experimental work and support the signal-to-noise hypothesis (Flach, Hagen, & Larish, 1992). The second environment involved collision avoidance. The key independent variables were speed of approach and the climb dynamics. The results showed that subjects were sensitive to both the dynamic constraints and uncertainties associated with action. Performance curves approached the optimal performance boundaries in state space. The buffer between the optimal boundary and the performance curves was proportional to the variability of responses. These two studies illustrate an active psychophysics paradigm that focuses on perceptual-motor coordination within closed-loop control tasks. The results are consistent with the logic of optimal control models that incorporate both dynamic constraints and uncertainty (perceptual and motor noise) as critical components in the model of the human operator.
Available at: http://works.bepress.com/john_flach/25/