MetaCartSign in to MyCiteSeer

Include Citations | Advanced Search | Help

Include Citations | Advanced Search | Help

  Design Patterns for Evolutionary Robotics

Download:
Download as a PDF
by Esben H. Ostergaard
http://www.adaptronics.dk/Publications/pdf/VikingPLoP2002.pdf
Add To MetaCart

Abstract:

The construction of controllers for robots performing simple tasks in wellknown and well-structured environments is reasonably well understood, and well established theories and communities exist. On the other hand, construction of controllers for robots placed in dynamic and unstructured environments can be very challenging, and no good general methodologies exist. Often, finding good solutions relies on the experience and intuition of the system engineer. Evolutionary robotics offers an approach to automate this process. However, designing an evolutionary robotic system is a non-trivial process, that requires perhaps even more experience and intuition than designing the controller itself. The development of a toolbox in the form of a design pattern language might significantly reduce the problems related to designing such a system. In this paper, four design patterns regarding evolutionary robotics is presented that could become part of design pattern language for evolutionary robotics. The four presented patterns focus on the principles and dynamics of evolutioanry robotics, rather than on implementation aspects. The first pattern can be regarded as a context for the subsequent three patterns.

Citations

1316 Genetic Algorithms + Data Structures = Evolution Programs. AI Series – Michalewicz - 1992
239 Understanding intelligence – Pfeifer, Scheier - 1999
184 Neural networks and fuzzy systems. A dynamical systems approach to machine intelligence. Englewood Cliffs, N.J – Kosko - 1992
115 Evolving mobile robots in simulated and real environments – Miglino, Lund, et al.
89 Noise and the reality gap: the use of simulation in evolutionary robotics – Jakobi, Husbands, et al. - 1995
17 Learning complex robot behaviours by evolutionary computing with task decomposition – Lee, Hallam, et al. - 1998
10 Sufficient neurocontrollers can be surprisingly simple – LUND, HALLAM - 1996
8 Reducing human design and increasing adaptability in evolutionary robotics – Floreano - 1997
7 Artificial evolution in the physical world – Thompson - 1997
4 Running across the reality gap: Octopod locomotion evolved in a minimal simulation – Jacobi - 1468
2 Evolutionary Robotics; Where From and Where To. Evolutionary Robotics: From Intelligent Robots to – Brooks - 1997
2 Issues in evolutioanry robotics – Harvey, Cliff, et al. - 1992