Evolving Controllers for Physical Multilegged Robots (2011)
Author: Vinod Valsalam
The videos linked below demonstrate the gaits produced by neural network controllers designed using the ENSO neuroevolution method for a physical quadruped robot. These controllers were evaluated in a physical simulation of the robot for walking on flat ground (1) when all four legs of the robot are functional and (2) when one leg is disabled to simulate a real-world motor failure. The controllers produced in the first experiment were evaluated further for generalization by reducing the maximum speed of the motors and by initializing one of the legs with a large error. In each experiment, generalization was also tested by placing the robot on different surfaces. These experiments show that the evolved controllers generalize well and are more robust against faults than a hand-designed PID controller, demonstrating the potential of the ENSO approach for real-world applications.

Demo website
Risto Miikkulainen Faculty risto [at] cs utexas edu
Vinod Valsalam Ph.D. Alumni vkv [at] alumni utexas net
Constructing Controllers for Physical Multilegged Robots using the ENSO Neuroevolution Approach 2012
Vinod K. Valsalam, Jonathan Hiller, Robert MacCurdy, Hod Lipson and Risto Miikkulainen, Evolutionary Intelligence, Vol. 5, 1 (2012), pp. 1--12.
Evolving Symmetry for Modular System Design 2011
Vinod K. Valsalam and Risto Miikkulainen, IEEE Transactions on Evolutionary Computation, Vol. 15, 3 (2011), pp. 368--386.
Utilizing Symmetry in Evolutionary Design 2010
Vinod Valsalam, PhD Thesis, Department of Computer Sciences, The University of Texas at Austin. Technical Report AI-10-04.
ENSO This package contains software implementing the ENSO approach for evolving symmetric modular neural networks. It also in... 2010