Explorations in Evolutionary Robotics

Abstract

We discuss the methodological foundations for our work on the development of cognitive architectures, or control systems, for situated autonomous agents. Our focus is the problems of developing sensorimotor control systems for mobile robots, but we also discuss the applicability of our approach to the study of biological systems. We argue that, for agents required to exhibit sophisticated interactions with their environments, complex sensorimotor processing is necessary, and the design, by hand, of control systems capable of such processing is likely to become prohibitively difficult as complexity increases. We propose an automatic design process involving artificial evolution, wherein the basic building blocks used for evolving cognitive architectures are noise-tolerant dynamical neural networks. These networks may be recurrent and should operate in real time. The evolution should be incremental, using an extended and modified version of a genetic algorithm.Practical constraints suggest that initial architecture evaluations should be done largely in simulation. To support our claims and proposals, we summarize results from some preliminary simulation experiments in which visually guided robots are evolved to operate in simple environments. Significantly, our results demonstrate that robust visually guided control systems evolve from evaluation functions that do not explicitly require monitoring visual input. We outline the difficulties involved in continuing with simulations and conclude by describing specialized visuorobotic equipment, designed to eliminate the need for simulated sensors and actuators.

A Geometry-Based Local Descriptor for Range Data

“…If explicit 3D shape descriptors are available, the grasping and picking-up parts of this problem can be significantly reduced [4].…”

What is the dynamical hypothesis?

“…Examples: cognitive neuroscience (Amit 1989;Babloyantz & Lourenco 1994;Cohen 1992;Guckenheimer et al 1993;Mpitsos, in press;Skarda & Freeman 1987); psychophysics (Gregson 1995); perception (Bingham et al 1997;Grossberg & Rudd 1992;McClelland & Rumelhart 1981;; motor control (Bullock & Grossberg 1988;Saltzman 1995;Turvey 1990); developmental psychology ; cognitive psychology (Busemeyer & Townsend 1993;Grossberg & Gutowski 1987;Grossberg & Stone 1986;Leven & Levine 1996;Tabor et al 1996); situated robotics and autonomous agents research (Beer 1995b;Cliff et al 1993;Smithers 1994a); artificial intelligence ( Jaeger 1996;Pollack 1991); social psychology (Kaplowitz & Fink 1992;Vallacher & Nowak 1993); ecological psychology (Kugler et al 1980; van Gelder: The dynamical hypothesis BEHAVIORAL AND BRAIN SCIENCES (1998) Turvey & Carello 1995); synergetics (Haken & Stadler 1990); morphodynamics (Petitot 1985b;Thom 1983;Wildgen 1982). Port and van Gelder 1995 is a representative sampling of the dynamical approach.…”

Untitled

“…Some of the Machine learning techniques becomes inapplicable to the learning reactive behavior problem because they require more information than the problem constraints allow. In the last few years, new approaches that involve a form of simulated evolution have been proposed in order to build autonomous robots that can perform useful tasks in unstructured environments [9,10,33]. The main interest in this approach is due to the dissatisfaction with traditional robotic and Artificial Intelligence approaches, and their belief that interesting robots may be too difficult to design.…”