Newsgroups: comp.robotics
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!gatech!howland.reston.ans.net!torn!watserv2.uwaterloo.ca!watserv3.uwaterloo.ca!kingcong.uwaterloo.ca!rbgorbet
From: rbgorbet@kingcong.uwaterloo.ca (Rob Gorbet)
Subject: Re: Limbed 'Stick Insect' Robot
Message-ID: <D7wJun.JJp@watserv3.uwaterloo.ca>
Keywords: Robot
Sender: news@watserv3.uwaterloo.ca
Nntp-Posting-Host: kingcong.uwaterloo.ca
Organization: University of Waterloo
References:  <sdimaio.1.2FA10C48@eleceng.uct.ac.za>
Date: Mon, 1 May 1995 13:51:57 GMT
Lines: 20

In article <sdimaio.1.2FA10C48@eleceng.uct.ac.za>, sdimaio@eleceng.uct.ac.za writes:
> Learning Algorithm for a Six (or Four) Legged Stick Insect Robot.
> 
> Is there somebody out there who has successfuly created a 
> 'learning algorithm' for for a limbed robot?

I believe I remember seeing, possibly a few years back now, just such
an article.  It was probably in IEEE Spectrum, or perhaps IEEE Transactions
on Robotics and Automation, or the Journal of the Society of Robotics and
Automation.  I don't have it anymore, and I'm not about to go looking
for it, but I'm pretty sure it dealt with the MIT(?) insect "Ghengis".

From what I remember, it used some sort of learning algorithm, since they
just seemed to let it go on its own.  It fell down a bunch of times when it
tried to move two support legs at the same time, but it eventually "learned",
and ended up walking very much like an insect would...

Perhaps someone else has the reference?

Rob.
