RE: minimal maze solvers

Hi John and Everyone,

Not sure about "would not approve". I get obsessive about rules that are open to interpretation. Something that defines the hardware is O.K. as there is no question about whether you meet the spec. or not. The competition can get a bit boring after a while though as, similar to the Singapore schools, the mice become very similar and the times nearly identical.

Schools use PICAXE a lot. The PICAXE 08M is a 12F683 and would represent a real challenge. Highcrest School are currently working on a F1 racer that uses two of these but a maze solver using one is probably impossible. Derek and Jim's PICONE with two bigger PICAXEs solves mazes well but has more resources than John proposes. A defined hardware competition based on a single suitable PICAXE might be interesting but I don't know which device would be appropriate.

If there is going to be a "one make series", please use something with a cheap (or free) development environment such as PICAXE but please let's not have complex rules about construction like Superbike, Moto GP or F1. Scrutineering is not a lot of fun!

Regards,

Duncan

-----Original Message-----
From: owner-micromouse@cs.rhul.ac.uk [mailto:owner-micromouse@cs.rhul.ac.uk]On Behalf Of WJP001@aol.com
Sent: 05 April 2009 22:10
To: micromouse@cs.rhul.ac.uk
Subject: Re: minimal maze solvers

David

    My definition is a very simple mouse which can achieve a time, albeit a slow one, under competition conditions. My mouse uses a PIC12F615. This device has 5 I/O pins and 1 input only pin. 1K words of flash programme memory and 64 bytes of RAM. The dynamic code has been tried on a small 4x3 maze and performs well but is yet to be tried on a full size maze. On a simulator, the navigation / maze solving code currently achieves a time with about 80% of mazes. By this I mean it finds it's way to the centre, knows how it got there, and is able to return to the start cell by an optimised route and finally sprints back to the centre using the same route, all in under the 10 minutes allowed by the rules. It is surprising how often it's chosen route is that which would be found by flooding. When dynamics are taken into account the it's optimised route is occasionally the better.

    I like the idea of a competition for minimal maze solving mice since the costs in both time and money are minimal but the intellectual challenge is worthwhile. Compared with a serious mouse it is rather like doing the Times crossword over morning coffee as opposed to designing the software to fly the Eurofighter. I demonstrated an early version of the navigation / maze solving code at Minos a few years ago but since no one else seemed interested in a competition for minimal mice, mine spends most of it's time gathering dust. The original target device was the PIC12C672 but since this is going slowly obsolete and making the mouse work with all of 128 bytes of RAM was no longer a challenge, I updated to the PIC12F615 when I last blew the dust off.

    The idea for the competition was a timed run with the time modified by a handicap dependant on the complexity of the hardware. So a mouse using a PIC10F2XX to produce a very slow time will beat something travelling at supersonic speeds with a Pentium. But do not ask me to think up a handicap system, my last effort was not well received.

    Any thoughts on a minimal maze solving mouse competition?

cheers

John

PS I had partaken of a certain amount of alcohol before replying to the emails re TIC and TID and apologise profusely for a rather feeble addition to the debate. ( Should have been Tid and Tego.)

PPS Microchip have supported micromouse in the past, they might support a "one make series" based on one of their devices such as the PIC12F615, this could go international, but Duncan would not approve!

<<
Email has been scanned for viruses by Altman Technologies' email management service
>>

Email has been scanned for viruses by Altman Technologies' email management service