Path: utzoo!attcan!uunet!seismo!sundc!pitstop!sun!pepper!cmcmanis
From: cmcmanis%pepper@Sun.COM (Chuck McManis)
Newsgroups: comp.sys.amiga
Subject: Re: Senior Project......
Message-ID: <77896@sun.uucp>
Date: 16 Nov 88 19:38:22 GMT
References: <5435@louie.udel.EDU>
Sender: news@sun.uucp
Reply-To: cmcmanis@sun.UUCP (Chuck McManis)
Organization: Sun Microsystems, Mountain View
Lines: 57

[Bitnet mailers are made out of rubber, everytime I throw a message at them
 it bounces right back to me.]

In article <5435@louie.udel.EDU> (Binoy James, Lehigh University.) writes:
> Now any brilliant suggestions that would be fun.  Any wacky, innovative,
> or stupid ideas are welcome.  I would be interested in doing something
> maybe robot-based, maybe with a tinkling of AI. Am I jumping too high or
> what? I don't know.  Let me know. I want some interesting ideas here.

Well one of the many projects that sits on my whiteboard at home and 
hasn't yet been done, is a robot simulator. Everyone who has built a
robot knows that it quickly becomes very expensive, and what you really
want to do is make it do something intelligent not just roll forward 
about 5 feet (+/- 1' :-)) and stop. 

Design notes (no code exists yet!) :

The simulator runs in a multiroom "universe". Rooms are 1000 X 1000 pixels
wide maximum (1 pixel == .5") Giving a maximum room size of about 41'
on a side. One can simulate most houses in this environment. This also
lets you use a superbitmap window to show the current position of the
robot. 

Robots will typically be about 24 - 50 pixels on a side (12" - 25") and
can be equipped with any equipment that can be modelled.

Sensors that are modelled are InfraRed, Ultrasonic, Bumpers, Radio Frequency
Beacons, and light detectors. 

The robot itself should have reasonable motors on it, I have a model for
some brevel motors (in my current robot) that travel at 30'/minute. 
Modeling such things as "carpet slip" is desirable.

The goals :
	o Build a simulator with a database for a 3 bedroom 1 1/2 bath
	  house available. 

	o Build the routines that allow the Robot to move around in this
	  simulator and to measure it's environment.

	o Design a simple interpretive language for the Robot that will
	  allow it to make moves and test input.

	o Design a dead reckoning program that will move the robot from
	  Point A to Point B.

	o Design a navigating program that uses sensor input to move the
	  robot from Point A to Point B
	
	o Build the simulated robot and compare it to the simulated 
	  performance. (optional)

Sound like a years worth?

--Chuck McManis
uucp: {anywhere}!sun!cmcmanis   BIX: cmcmanis  ARPAnet: cmcmanis@sun.com
These opinions are my own and no one elses, but you knew that didn't you.