Paper on connectionist robot control

[Lisa Meeden]

The following paper is available in the neuroprose archive.  It will
also appear in the "Proceedings of the Fifteenth Annual Conference of
the Cognitive Science Society".  Unfortunately for those unable to use
FTP, we cannot provide hard copies.

	Emergent Control and Planning in an Autonomous Vehicle

	   Lisa Meeden*      Gary McGraw*+      Doug Blank*
		   *Department of Computer Science
	    +Center for Research on Concepts and Cognition
			  Indiana University

			       Abstract

We use a connectionist network trained with reinforcement to control
both an autonomous robot vehicle and a simulated robot.  We show that
given appopriate sensory data and architectural structutre, a network
can learn to control the robot for a simple navigation problem.  We
then investigate a more complex, goal-based problem and examine the
plan-like behavior that emerges.


To obtain a copy:

	ftp cheops.cis.ohio-state.edu
	login: anonymous
	password: <your email address>
	cd pub/neuroprose
	binary
	get meeden.robot.ps.Z
	bye

Then at your system:

	uncompress meeden.robot.ps.Z
	lpr -P<printer-name> meeden.robot.ps