From Animals to Animats 7,
The seventh international conference on the Simulation of Adaptive Behavior
(SAB’02)
Workshop on Motor Control in
Humans and Robots: on the interplay of real brains and artificial devices
A.Karniel*1; K.M.Fleming1;
V.Sanguineti2; S.T.Alford3; F.A.Mussa-Ivaldi1
1. Department of Physiology, Northwestern University Medical School,
Chicago, IL, USA; 2. Dipartimento di Informatica Sistemistica e Telematica,
Universita' di Genova, Genova, Italy; 3. Department of Biological Sciences,
University of Illinois at Chicago, Chicago, IL, USA
In order to study the neurobiological mechanisms of learning, we developed
a research tool that includes the brainstem of a lamprey and a two-wheeled
robot interconnected in a closed loop. Two electrodes applied stimulations to
the neural tissue. The stimulation frequency was set to be proportional to the
light intensity measured by sensors on the right and left sides of the mobile
robot. The velocity commands to the
right and left wheels were proportional to the population spike rates recorded
by two recording electrodes. In most
cases, the robot moved approximately in the direction of the source of light.
We fitted ten two-input/two-output neural network models and observed
the generalization error of each model.
We found that a dynamic model was significantly superior to a static
model even when the number of parameters was smaller. Additional findings led
us to conclude that the main origin of this dynamic behavior is local
ipsilateral influence of the previous state on the current state. An analysis of the model that fits adaptive
behavior of the preparation showed a significant change in one of the recurrent
connections. We discuss these results
in the context of possible cellular mechanisms that might explain the behavior
of this neuro-robotic preparation.