Papers Available: Neuro-Evolution in Robotics

[moriarty@AIC.NRL.Navy.Mil]

The following two papers on applying neuro-evolution to robot arm
control are available from our WWW page:

http://www.cs.utexas.edu/users/nn/

Source code for the SANE system is also avaiable from the WWW site.

------------------------------------------------------------------------

        Evolving Obstacle Avoidance Behavior in a Robot Arm

            David E. Moriarty and Risto Miikkulainen

To Appear at From Animals to Animats The Fourth International
Conference on Simulation of Adaptive Behavior (SAB96).  Cape Cod,
MA. 1996

8 pages 

Abatract:

Existing approaches for learning to control a robot arm rely on
supervised methods where correct behavior is explicitly given.  It is
difficult to learn to avoid obstacles using such methods, however,
because examples of obstacle avoidance behavior are hard to generate.
This paper presents an alternative approach that evolves neural
network controllers through genetic algorithms.  No input/output
examples are necessary, since neuro-evolution learns from a single
performance measurement over the entire task of grasping an object.
The approach is tested in a simulation of the OSCAR-6 robot arm which
receives both visual and sensory input.  Neural networks evolved to
effectively avoid obstacles at various locations to reach random
target locations.

------------------------------------------------------------------------

            Hierarchical Evolution of Neural Networks

            David E. Moriarty and Risto Miikkulainen

Technical Report #AI96-242, Department of Computer Sciences, The
University of Texas at Austin.

16 pages

Abstract:

In most applications of neuro-evolution, each individual in the
population represents a complete neural network.  Recent work on the
SANE system, however, has demonstrated that evolving individual
neurons often produces a more efficient genetic search.  This paper
explores the merits of neuro-evolution both at the neuron level and at
the network level.  While SANE can solve easy tasks in just a few
generations, in tasks that require high precision, its progress often
stalls and is exceeded by a standard, network-level evolution.  In
this paper, a new approach called Hierarchical SANE is presented that
combines the advantages of both approaches by integrating two levels
of evolution in a single framework.  Hierarchical SANE couples the
early explorative quality of SANE's neuron-level search with the late
exploitative quality of a more standard network-level evolution.  In a
sophisticated robot arm manipulation task, Hierarchical SANE
significantly outperformed both SANE and a standard, network-level
neuro-evolution approach, suggesting that it can more efficiently
solve a broad range of tasks.

------------------------------------------------------------------------


Dave Moriarty

Artificial Intelligence Laboratory      
Department of Computer Sciences 
The University of Texas at Austin       
moriarty at cs.utexas.edu  
http://www.cs.utexas.edu/users/moriarty
http://www.cs.utexas.edu/users/nn