Behavior-Based Robotics


Introduction

Behavior-based robotics or behavioral robotics or behavioural robotics is the branch of robotics that does not use an internal model of the environment. For instance, there is no programming in the robot of what a chair looks like, or what kind of surface the robot is moving on - all the information is gleaned from the input of the robot's sensors. The robot uses that information to react to the changes in its environment.

Behavior-based robots (BBR) usually show more biological-appearing actions than their computing-intensive counterparts, which are very deliberate in their actions. A BBR often makes mistakes, repeats actions, and appears confused, but can also show the anthropomorphic quality of tenacity. Comparisons between BBRs and insects are frequent because of these actions. BBRs are examples of Weak artificial intelligence.

The school of behavior-based robots owes much to work undertaken in the 1980s at the Massachusetts Institute of Technology by Professor Rodney Brooks, who with students and colleagues built a series of wheeled and legged robots utilising the subsumption architecture. Brooks' papers, often written with lighthearted titles such as "Planning is just a way of avoiding figuring out what to do next", the anthropomorphic qualities of his robots, and the relatively low cost of developing such robots, popularised the behavior-based approach.

Brooks' work builds - whether by accident or not - on two prior milestones in the behavior-based approach. In the 1950s, Walter Grey Walter, an English scientist with a background in neurological research, built a pair of vacuum tube-based robots in the 1950's, which were exhibited at the Festival of Britain in 1951, and which have simple but effective behavior-based control systems.

The second milestone is Valentino Braitenberg's 1984 book, "Vehicles - Experiments in Synthetic Psychology" (MIT Press). He describes a series of thought experiments demonstrating how simply wired sensor/motor connections can result in some complex-appearing behaviors such as fear and love.

Some of the latest work in BBR is from the BEAM robotics community, which has built upon the work of Mark Tilden. Tilden was inspired by the reduction in the computational power needed for walking mechanisms from Brooks' experiments (which used one microcontroller for each leg), and further reduced the computational requirements to that of logic chips, transistor-based electronics, and analog circuit design. [Wikipedia]

Evolutionary Robotics 


Introduction

Evolutionary Robotics is a new technique for the automatic creation of autonomous robots. It is inspired upon the Darwinian principle of selective reproduction of the fittest. It is a new approach that looks at robots as autonomous artificial organisms that develop their own skills in close interaction with the environment without human intervention. Heavily drawing from natural sciences like biology and ethology, evolutionary robotics makes use of tools like neural networks, genetic algorithms, dynamic systems, and bio-morphic engineering. The resulting robots share with simple biological systems the characteristics of robustness, simplicity, small size, flexibility, and modularity. In evolutionary robotics, an initial population of artificial chromosomes, each encoding the control system of a robot, is randomly created and put into the environment. Each robot is then free to act (move, look around, manipulate) according to its genetically specified controller while its performance on various tasks is automatically evaluated. The fittest robots then "reproduce" by swapping parts of their genetic material with small random mutations. The process is repeated until the "birth" of a robot that satisfies the performance criteria.

EvoROB (EvoNet -European Network of Excellence in Evolutionary Computing- working group in evolutionary robotics) definition for Evolutionary Robotics:

"Evolutionary robotics uses simulated evolution to automatically design autonomous robots. It takes a bottom-up approach to     robotic learning, synthesising controllers through close interaction with the environment rather than through human design."

White Papers

  • Beer R. D. & Gallagher J. C. (1992) Evolving dynamical neural networks for adaptive behavior. Adaptive Behavior, 1:91-122.
  • Brooks R. A. (1989) A robot that walks: Emergent behaviors from a carefully evolved network. Neural Computation, 1:253-262.
  • Chavas J., Corne C., Horvai P., Kodjabachian J. & Meyer, J-A. (1998) Incremental evolution of neural controllers for robust obstacle-avoidance in Khepera. In P. Husbands & J-A. Meyer (Eds.), Evolutionary Robotics. First European Workshop, p.227-247. Berlin: Springer-Verlag.
  • Floreano D. (1997) Reducing human design and increasing adaptability in evolutionary robotics. In T. Gomi (Ed.), Evolutionary Robotics: From Intelligent Robots to Artificial Life. Ontario, Canada: AAI Books.
  • Nolfi S. (1996). Adaptation as a more powerful tool than decomposition and integration. In T. Fogarty & G. Venturini (Eds), Proceedings of the Workshop on Evolutionary Computing and Machine Learning, 13th International Conference on Machine Learning, Bari.

Free Software

Books

  • Evolutionary Robotics
  • From Animals to Animats
Content is available under GNU Free Documentation License 1.2
Copyright © 2008 irobotics.org