Now we are talking ...steam powered mice! I also had the idea to do this, but Martin put me off by telling me about these damned rules!! However .. we can still do it, duncan, but we'd need to use compressed air/CO2, not combustible fuel. Anybody can tell me what size cylinder, at what pressure, we'd need to move a half-kilo round a maze for 10 minutes? Tony -----Original Message----- From: owner-micromouse@cs.rhul.ac.uk on behalf of Duncan Louttit Sent: Sun 7/4/2004 12:21 PM To: micromouse@cs.rhul.ac.uk Cc: Subject: RE: Rules or specifications? Hi Everyone, There are other physical limits imposed on the mouse, e.g. must not pollute, must not leave a trail. . . These rules seem to be necessary, but should be reviewed periodically. I would still like to make a steam powered mouse sometime! The point about the slots in the walls is that they DO cause problems. I have seem several mice caught by them and I have had problems with sensor systems that "see" the slots. It is no good building a maze with 3mm slots when the rules specify 1mm. This is what I mean by the rules specifying the actual competition rather than some idealised version. If we cannot build a maze with a maximum iregularity in the floor of 1mm, we should specify something larger. It is not fair for a mouse to fail because the maze does not meet its specification. The problem with "non contact" is that "sensors" can be specified simply, but it is not obvious which part of a DIM type machine is the "sensor". Some mice with optical sensors still rely on the shape of the mouse to recover from contacts with the wall and it is not unknown for a mouse with optical sensors (e.g. DOT) to "lean" on the wall for maximum straight line speed. A few collapsible walls placed at the judges discretion will solve the problem. The school's formula 3 does effectively that: "The judges may put obstacles such as cotton reels with their inside edges no less than 13 centimetres from the edge of the line (inside and outside the course) at points that they choose. If a mouse moves one of the obstacles, that lap cannot be used for scoring." I like emails as a way to produce a set of rules. It is not always possible for everyone to get to a meeting, whereas all mousers that I know of have email. This discussion could go on a long time. Perhaps we should start going through the rules one clause at a time to identify where the existing rules are O.K. so that those parts can be frozen, and agree where there are differences of opinion. Where there are differences, someone will have to make a decision but whoever builds next year's maze should be aware of the tolerances that they will have to deliver on the day and whoever judges next year's competition should be prepared for appeals regarding whether a particular mouse is "contact" or not. Regards, Duncan At 00:09 04/07/04, you wrote: Duncan et al.... Some of this is in response to previous postings, so anyone that is confused look back through the past weeks worth! "Make the fastest maze solver" is clearly the sole AIM of anyone who wants to win the senior competition. It could never be classified as a rule. The specification of the mouse is almost entirely up to the mouse builder- the ONLY physical limits placed on it by the rules it must fit inside the maze walls and not damage them when it moves around. The mouse just has to meet the requirement to have a total size of no more than 25cm by 25cm max above the walls, and it must clearly fit inside the maze walls. In this case, this is an adequate specification. The MAZE construction obviously needs to be specified, with stated tolerances, such that the mouse-builder can maximise his/her effectiveness in a maze that conforms to these agreed specifications. For the SENIOR competition, I don't think we need to put much more effort (if any) into the specification - the mazes seem to work well enough, and I have not heard a great deal of complaint about the grooves in pegs, or gaps in walls - admittedly I have only attended the last three events. Since the maze is used in the same form worldwide, we would be silly to change anything. The way the mouse is 'scored' is determined by the RULES. Here again, for the Senior competition there is not a lot to do. We have already had the suggestion that we adopt the (in my opinion) slightly more evolved APEC rules - and I really can't see why not. I make the distinction re: 'rules' and 'specification', because I think it helps to clarify those issues that are to do with the way that the competition is run i.e rules, and those issues that that are to do with the design of the competition arena. We can then concentrate on addressing small problems, and we should reach consensus quicker. "The rules must describe the competition as it is, warts and all, not as we would like it to be.", Duncan Loutit, 07/03/04. Sorry Duncan, can't agree there ..... surely the point of these discussions is to get rid of a few of the warts? I take your (Duncan's) point about the changing needs of the competition. We spent a lot of time deliberating whether to make a distinction between contact and non-contact wall followers, in the run-up to MM2004. My interpretation of contact was a device that navigated purely by contact means - sensors or not. Non-contact was a device that navigated by non-contact sensors. The occasional brush with the wall should not mean a transfer from the non-contact class to the contact class. We made the attempt - I think it was worth it - and I think the 3-D follower competition has to evolve as an appropriate event for schools and Universities. There is some work to be done here, but not a great deal - the classification of a device as contact/non-contact should not be a difficult matter. Having walls that fall over really doesn't solve anything unless they are ALL capable of falling over - mind you, we could then combine the event with the world domino toppling championships. An aside: Solo Domino Toppling The greatest number of dominoes set up and toppled single-handedly is 281,581 out of 320,236 by Klaus Friedrich at Fürth, Germany, on January 27, 1984. The dominoes took 31 days to set up and just 12 minutes 57.3 seconds to fall. Perhaps we could race a wall-follower against the toppling line of dominos? I could doubtless think up a number of possibilities for an intermediate competition, based on 3-D mazes. e.g. 1. Reduce the size of the maze to, for example, 8x8 cells. This should increase the throughput, and make the competition more exciting - we could have 4 competitiors going at once, in identical mazes. 2. Formulate the maze as a circuit, whereby two mice start at opposite sides, similar to time trials in cycling. 3. Run a 3-D rat-race, using the 4 maze segments to make a 6m long course with 4 lanes - purely for wall-following ability and speed. Doubtless all thought of before - before you all rush to tell me - but the time is right for us (at tic) now. In the past I would have said that some of these were a little too simple for undergraduates - but with the change in emphasis on practical work over the past few years(for the worse), we would consider ourselves lucky to have a student that can build a maze solver. We had one this year, who had problems with the lighting (not using ambient cancellation). We could have more as we continue the development of our hardware platforms, and if my Engineering Clubs take off. In reality, a wall-follower is all that we can target at the moment - and so I have a need to develop 1-3 above. There is clearly an overlap with the Schools competition, but we are talking about 3-D routes. To summarise: It is my view that the Senior competition should be left pretty much as it is, with the slight mods mentioned at the start of this expansive document. It is a damned tricky event, but we have new-blood coming in, and it would be a shame to change for the sake of change before they have had a chance to strut their stuff. We need to work on the intermediate competitions. Email discussions such as this are no substitute for a full-blooded open debate. We have to get together in one place, with the sole aim of focussing what we do. Why don't we organise a one-day event, in early September, and keep the iron hot? If you want, I can set it up at TIC - I have no preference. Regards to all Tony -----Original Message----- From: owner-micromouse@cs.rhul.ac.uk on behalf of Duncan Louttit Sent: Sat 7/3/2004 4:15 PM To: micromouse@cs.rhul.ac.uk Cc: Subject: RE: Rules or specifications? Hi Everyone, I couldn't get to my emails for a couple of days and wow there's a lot of stuff about micromouse! Great! I cannot tell the difference between rules and specifications. The abstract concept "make the fastest maze solver" might qualify as a rule under Tony's ideas but any real mouse has to have a specification (as he puts it) for the maze, any limitations on the mouse (size etc.), "solving", and the way it will be scored. Any change to any of these specifications, changes the nature of the competition. I entered two mice in the "non-contact" wall follower class partly because I was not sure how "non contact" would be interpreted. One definitely never touched the walls and came (I think) last. One brushed them occasionally and came second. "The rules" must completely specify the competition. Changes to the rules imply changes to the mice (consider the confusion as to which of Dave Otten's mice won). The rules must be clear, unambigous and complete and they must be defined many months before competition date. The rules must describe the competition as it is, warts and all, not as we would like it to be. I really like the idea of collapsing walls. That is the kind of rule that is easy to judge and clear to understand. In the schools competition there is a similar problem for deciding whether the mouse had hit the restraining wall at the end of the rat run. The solution was to make the wall out of expanded polystyrene. If the mouse moved the wall it suffered the 5 second penalty. Regards, Duncan
<<winmail.dat>>