Evolutionary robotics: From behaviorism to embodied cognition

Abstract

Over the past decade we have witnessed the emergence of a new paradigm in control systems development, evolutionary robotics, in which robotic control systems are evolved rather than constructed by hand. Robotic hardware designs are becoming more complex as the variety and number of on-board sensors increases, and as greater computational power is provided onboard in ever-smaller packages. Advances in hardware, however, do not automatically translate into better software for controlling complex robots. Evolutionary techniques hold the potential to solve many difficult problems in robotics which defy simple conventional approaches, but present many challenges as well. Numerous disciplines including artificial life, cognitive science and neural networks, rule-based systems, behavior-based control, genetic algorithms and genetic programming have contributed to shaping the current state of evolutionary robotics. This paper provides an overview of developments in the emerging field of evolutionary robotics, and discusses some of the opportunities and challenges which currently face practitioners in the field.