Meeting Abstract

P3.88  Jan. 6  Chasing what you can(not) see: Robotic tests of tiger beetle visual pursuit strategies MAPES JR, NJ*; BELANGER, JH; Louisiana State University, Baton Rouge jimb@lsu.edu

Visual predators may detect prey items while moving, stationary, or both. Information about the location and velocity of a moving target can be degraded both by motion of the target and of the pursuer, yet many animals solve this problem very well. Gilbert (1997, J Comp Physiol A 181: 217) performed a series of behavioral experiments with tiger beetles, followed by computer simulations, to determine possible pursuit strategies. He suggested that tiger beetles use a closed-loop, position-sensitive tracking system to pursue prey, provided that image contrast and background complexity allowed the beetle to continuously detect the prey. Otherwise, the beetles use an open-loop, iterative strategy in which they alternately chase prey, stop long enough to relocalize it, and then pursue it again. Gilbert suggested that the switch between these strategies may be influenced strongly by limitations in signal detection, particularly spatio-temporal contrast. Contrast should be dependent on both the visual properties of the target, and the background against which it is being viewed. To examine the complex interactions between these visual properties, and those of the putative control system, we have used a small mobile robot (Khepera) equipped with a video camera system, to model tiger beetle pursuit behavior. We matched as closely as possible the properties of the video system to those of the tiger beetle visual system, and challenged the robot to pursue visual targets in a behavioral arena. Using this �robot beetle�, we are replicating the insect behavior experiments, but in a system where we have complete knowledge of both the properties and the performance of the control systems. Presently, we are using the system to determine the level of environmental complexity (low spatio-temporal contrast) necessitating a switch from closed-loop to open-loop pursuit.