Track Design

[HarryBotter]

Let's use this thread to document the characteristics of exisiting RC model race tracks. Try to keep posts to this thread on topic and about course characteristics and design.

Harry

[dporter]

I've been following the general thread on this topic and I think the key is to have a track in mind first and then many of the other pieces will follow. I agree with some of the other posts that this is something that will grow in stages, so the first track should be basic in design. I also think that "competing with R/C cars" should be the ultimate goal.

The first track should be similar to indoor R/C courses. The popular track in my area is carpeted with 4 X 4 wood borders. The track uses movable borders so that the layout can be easily changed.

I would suggest that a simple oval track be used to begin with. The width and length could be standardized but this may be best left open. As far as any other reference points I would suggest only a start/finish line and the standard 4 X 4 borders. This would allow a wide variety of methods for negotiating the track without forcing tracks to change their layouts and add IR emitters or other devices.

While this may seem difficult at first, I have seen mindstorms RCX robots using just touch sensors and dead reckoning to negotiate similar track designs.

[hgordon]

The choice of sensors defines some of the track characteristics.

At the most primitive level, you could build a functional vehicle with nothing more than IR proximity sensors for wall or line following with obstacle avoidance.

The next level of sensors would be cameras and/or laser range finders, and you likewise need continuous "walls" or lines that will register with the sensors.

In either case, this requires a minimum-height continuous wall or some continuous lines.

By the way, we should assume that these races have a referee who deals with cars that go off-track, get flipped, travel the wrong way, get stuck, etc.

[dporter]

In either case, this requires a minimum-height continuous wall or some continuous lines.

By the way, we should assume that these races have a referee who deals with cars that go off-track, get flipped, travel the wrong way, get stuck, etc.


I agree, I think the key is to adapt the cars to current tracks and their 3-4" border walls. While there would be a number of ways of racing by changing the tracks, I think to meet the overall goal of competing with R/C cars we are going to be limited to what R/C racers currently use.

This would probably go in a different thread but I would also imagine that some sort of emergency shut-off would be needed just in case a car did spin or flip.

Thanks
Dporter

[HarryBotter]

Track configuration: Since I've seen the term 4 x 4 mentioned a few times why don't we just settle on track boundary walls 3 1/2 inches high and specify that they be painted white for increased reflectivity.

Harry

[hgordon]

How about some guidelines on overall track size (outside dimensions, e.g. 12-ft x 8-ft) and width of the track surface (min/max distance between inside and outside walls, eg 2-ft) ?

[HarryBotter]

Let's hope we get some indication of these details from the experienced RC racers. We would need a handle on number of vehicles per race and track width to accomodate that number.

Are you actually thinking of building a full sized track? I'm just going to put together a representative section that will allow sensor and algorithm testing. Quite frankly, given my time commitments right now I'll be moving rather slowly anyway.

Harry

[dporter]

I'll probably do the same thing Harry. I'll set up just a section at first and then if I have to head out to the track in town on an open track night to give things a try on a regular size track. Besides if I have somethng that works at the track I might drum up some interest.

Doug

[hgordon]

Is this what you guys have in mind for a full-size track ? -

http://www.hpiracing.com/walkthroughs/microtrack.html

http://www.hpiracing.com/graphics/misc/microtrack/DSCN0979k.jpg

[dporter]

That's nice. Something like that would be about the right size. It looks like it can also be reconfigured to other shapes.

[TheDuck]

I think adapting to the existing guidelines for racing as much as possible would be great. Imagine showing up at a regular event. "Hey, where's your R/C radio?" ;) Tell me that wouldn't be some effective marketing!

[HarryBotter]

Looking at this site suggests that the minimum track size we should be looking at is probably 8' x 20'.

http://brpracing.com/18scale.html

Harry

[PeterD]

Hi Harry,

I agree that 8' x 20' would be an absolute minimum.

The BRP Racing cars are intended for oval track racing. If the track is to have both lefthand and righthand turns, then the area will have to be larger. The sample track on the HPI site is 16' x 32'.

I think it would be valuable to spec some minimum dimensions for the different features of the track, rather than outside size. For example: minimum turn radius - 4', minimum track width - 3' 6". If track designers have guidelines for these features then they can use their imaginations to build a layout that fits the space available.

Looking at this site suggests that the minimum track size we should be looking at is probably 8' x 20'.

http://brpracing.com/18scale.html

Harry

[hgordon]

With those dimensions, you could fit a "figure-8" in an 8-ft wide track. Anyone for Demolition Derby ?

As Harry originally proposed, an oval track sounds like a good plan for testing and early racing. We can reverse the direction of racing to verify that left-turn algorithms work as well as right-turn algorithms.

[wrighthobbies]

Is there a way to design the track so the bots can determine they are going the right direction?

The width of a standard RC track would make it easier for a bot to turn itself around when trying to avoid other cars and the wall. It would make for fun watching but not bode well for the bots :)

[HarryBotter]

Is there a way to design the track so the bots can determine they are going the right direction?
Someone made the suggestion to color the track boundaries using different contrasting colors. Then by tracking these colors, correct orientation could be worked out.
The width of a standard RC track would make it easier for a bot to turn itself around when trying to avoid other cars and the wall. It would make for fun watching but not bode well for the bots :)
Fun to watch and should work with proper obstacle avoidance algorithms. I suspect the wise would employ some sort of contact sensor scheme as backup. :)

Harry

[hassenplug]

Another idea that we use here (in Indiana) is a somewhat fancy track. The track has a gradient, so it's black in the middle, and goes to white on the outside. It's an oval that fits on a 4x8 table.

That track makes it very easy to build a robot that can race. The robot just needs one light sensor, and it "knows" where it is on the track.

On the down side, a good track costs as much as a robot to race on it. But, if you have a track, anyone can easily build a robot.

Steve

[wrighthobbies]

Steve, do you have any pictures of that track? I'd like to see how that looks and works...

[hassenplug]

Steve, do you have any pictures of that track? I'd like to see how that looks and works...

We run this event every May. Here are the full rules...

http://www.teamhassenplug.org/LafLRC/Indy_5_Rules.pdf

There's a picture of the track in there. We've had someone print it out with a four foot wide printer, but paper is not very good.

Steve

[wrighthobbies]

Hmm, I wonder how else you could produce it with precise and consistent gradience...

Looks like a great way to do it, the bots pick a gradient value and follow it around the track. Heh, a flashlight could really screw up the race though :)

[hassenplug]

Looks like a great way to do it, the bots pick a gradient value and follow it around the track. Heh, a flashlight could really screw up the race though :)

It is fun!

A good robot will not turn too much on the straights, but will turn quickly in the corners.

It has one element we don't deal with in the Chibots events: head-to-head racing.

Here's a clip from one of our better races.

http://www.brickshelf.com/gallery/Hassenplug/Club/Indy506/rc1vscockroach.wmv

Steve

[dporter]

I like the track idea that you are using. You said that you have a meet in May in the Indianapolis area? Do you have different competition classes or beginners groups? I think my High School students would like something like this and Indy would be a reasonable trip for us.

Doug

[hgordon]

Steve -

That's very cool. Thanks for sharing the video !

[Kevin Berry]

I'm betting this would be a "simple" project for a decent graphics art/banner making outfit. They are just as expert at that as we are at building robots.

Uh, let me rephrase that .... time to call in a professional!

[wrighthobbies]

It is fun!

A good robot will not turn too much on the straights, but will turn quickly in the corners.

It has one element we don't deal with in the Chibots events: head-to-head racing.

Here's a clip from one of our better races.

http://www.brickshelf.com/gallery/Hassenplug/Club/Indy506/rc1vscockroach.wmv

Steve

True, the head to head racing looks like fun! I love the way the black bot overshoots the corners so much after it gets in the lead and still gets back on track.

Do you think the track would still work if the paper was laminated or made of some sort of plastic? The track would be more durable that way.

If a poster/sign shop printed it, you could go 4x12 or 4x16 too.

[hassenplug]

You said that you have a meet in May in the Indianapolis area? Do you have different competition classes or beginners groups? I think my High School students would like something like this and Indy would be a reasonable trip for us.


Actually, we live about an hour north, in Lafayette. We've done it three times so far, but we haven't had divisions, yet. But, I could see doing that in then future.

Last year, we had about 16 robots, which made for a really good half-day event.

I'm betting this would be a "simple" project for a decent graphics art/banner making outfit.

A couple years ago, I looked into getting one made by a banner company. I think they said it would be a couple hundred dollars. we didn't want to spend that, so we scrapped the idea.

However, it's worked out very well, so I may consider it, again.

True, the head to head racing looks like fun! I love the way the black bot overshoots the corners so much after it gets in the lead and still gets back on track.

Do you think the track would still work if the paper was laminated or made of some sort of plastic? The track would be more durable that way.

If a poster/sign shop printed it, you could go 4x12 or 4x16 too.

Yes, I think I'll look into doing it, again. One advantage of keeping it on a 4x8 area is that it will work on an FLL table. Then, you have walls to keep robots from driving off, and back on the track... :)

Steve