Challenge!

[TheDuck]

I would like to see a challenge item. Something based on real events might make it more interesting as well.

For example, there was a recent tragedy regarding a mine. There was a wheeled robot that didn't make it far past the mouth of the mine because it got stuck in mud. It was tracked but apparently that wasn't enough. Had it made it further it would have to contend with narrow passageways as well. What are some conceptual designs, or ones based on examples seen in the field today, that might have circumvented these challenges? Once on station, how do we identify where these people are?

Another example is the new FIRST competition; a fusion between soccer and basketball.

Who knows? Maybe the suggestions discussed on this site could be contenders in these competitions!

[GoRobotics.net]

So, you're thinking of a weekly/monthly/quarterly "design challenge" of sorts?

That's an interesting idea. Perhaps there could be prizes involved. The only problem is that coming up with good challenges each time might be difficult.

[Kevin Berry]

I think NASA does a pretty good job on their Centennial challenge series - ribbon climbers, excavators, etc.

[TheDuck]

Yes, I would have to agree. You would want them to be, not simple, but straightforward so as not to have people commiting an unreasonable amount of their lives to the challenge. It should also focus on a single task rather than trying to solve a raft of problems all at once.

The nice thing is that the magazine is currently quarterly so there is some time between challenges. If we were clever we could ensure that each challenge built upon the one before.

Now I think that tank is a fantastic start...

[ben-ja-men]

i think that a competition using a featured robotic base which everyone must use would save alot of time and make the playing field level. ppl could send video entrys along with details of programs etc. so that ppl who get stuck can learn via the process.

[TheDuck]

Nice idea. There are a few platforms out there to play with, too. Mass market such as VEX, Mindstorms, (soon to be) NXT and some more serious (read, "expensive") such as Khepera, Koala and so on.

However, it's tempting to make our own such as furthering the WiFi Tank or taking an off-the-shelf R/C car or truck and starting to add some smarts to that.

All in favor of the tank? :wink:

[ben-ja-men]

All in favor of the tank? :wink:

id tentatively say yes ive ordered the past issues and am eagerly awaiting :D

[donkren]

I'd like to register my vote for the Vex Starter Kit as a platform for Challanges.

It's available everywhere, very flexible (and HALF OFF right now.)

[HarryBotter]

I would like to see a challenge item

I'd like to revisit this idea, I think it has a lot of merit. :)

How about setting up a simple challenge that could be met by someone with just the basics, say a few pieces of liteply or acrylic, a couple of servos, perhaps a sensor or two and a basic stamp. These items could be used to build a line follower or an object finder, perhaps even a ribbon climber.

I'm sure ROBOT would give the event some coverage and perhaps even offer a free subscription as a prize.

Any ideas, comments? Perhaps a poll to gauge interest?

Cheers, Harry

[tatwood]

I like the idea of holding a challenge that is sponsored by Robot magazine with prizes that would include subscriptions or sub extensions, and perhaps even prizes from robot manufacturers. Keeping it simple and affordable would be the ticket for the largest participation. It probably makes the most sense to skip an issue between announcing and awarding prizes. If someone doesn't get an issue till 6 weeks into the on-sale period, getting the robot built and submitted in the remaining few weeks before the next issue is put to bed can paint the competitor and the magazine into a time corner.
I could see starting with something very simple--as Harry said, two servos or two motors, a Basic Stamp processor, and a simple task like a line tracking application. What is the absolute, least expensive, simplest design? I could see a later contest to see who can produce the fastest line tracker. Some are much faster than others. We could standardize the line track...
If the contest took off, there could eventually be multiple challenges. One using Basic for newcomers, and one using C. I like the idea because it could involve readers with the magazine--winning entries, or just plain interesting entries, could be posted on our website. Video, code, the works.
Anybody like this idea? Launching with a simplest line tracking design?

[HarryBotter]

I like the idea of holding a challenge that is sponsored by Robot magazine with prizes that would include subscriptions or sub extensions, and perhaps even prizes from robot manufacturers. Keeping it simple and affordable would be the ticket for the largest participation. It probably makes the most sense to skip an issue between announcing and awarding prizes. If someone doesn't get an issue till 6 weeks into the on-sale period, getting the robot built and submitted in the remaining few weeks before the next issue is put to bed can paint the competitor and the magazine into a time corner.
I could see starting with something very simple--as Harry said, two servos or two motors, a Basic Stamp processor, and a simple task like a line tracking application. What is the absolute, least expensive, simplest design? I could see a later contest to see who can produce the fastest line tracker. Some are much faster than others. We could standardize the line track...
If the contest took off, there could eventually be multiple challenges. One using Basic for newcomers, and one using C. I like the idea because it could involve readers with the magazine--winning entries, or just plain interesting entries, could be posted on our website. Video, code, the works.
Anybody like this idea? Launching with a simplest line tracking design?

Thanks for the magazine support! I'd think the first task would be to measure the amount of interest in this type of competition. Traffic in some of these forums gets quite low over the summer months so this may not be the best place to guage interest right now. How about a small notice in the next issue asking for feedback from interested parties?

And if you're reading this thread now, how about leaving a comment? Lets see if we can make this happen!

Cheers, Harry

[I D Miller]

I could totally dig for that let me see if i got what it could take to do it.

Video cam / Web Cam....................check
broad band internet.......................check
imigination....................................che ck
fear.............................................n ah
excitement........................oh yeah giddy like a school girl.


can we do it?
can we do it?
can we do it?

purty please with sugar on top with a cherry?


ok now for the adult side. I think that would be a great idea and the idea to skip an issue before posting results awesome so a bi anual contest. another small prize would be to meet some people who are into bots. just another idea. the only think that i would have a problem with is if there is a request to use certain componets. my vote is yes!

"pieces" be with you.
Dee

[Kevin Berry]

So how do you do a comp without people flying around the country? would these be "submit your design and a video" type things? Or "mentored" by a school administrator? Or would Robot pay for the best 5 to come to a comp?

[HarryBotter]

So how do you do a comp without people flying around the country? would these be "submit your design and a video" type things? Or "mentored" by a school administrator? Or would Robot pay for the best 5 to come to a comp?

Good question with two good suggestions and another interesting one. :)

Another might be a just a design contest, given a set of parameters. Like a line-follower or a bot that could follow a string between two points. A prize could be awarded based on submitted plans or pictures.

Or a classroom project in a science class. Submissions by a class could be sent to Robot magazine where the 'competition' of all submissions could be run and evaluated by staff and a winner(s) determined.

Obviously, face-to-face competition would initially be too costly but could eventually become reality with enough participation. Perhaps even a specific event category at the annual Robogames.

Harry

[tatwood]

We are considering starting the Challenge with our Spring issue, which will be available in late January. However, before we commit, we'd like more feedback, here, from interested roboticists. A lot of good ideas have been suggested; We'll put together a group of our most active contributors to sort out some ground rules, but, as I mentioned, we are still looking for more input and suggestions.

Initially, the challenge could be judged by a group of editors and magazine contributors, and if manufacturers participate, some might be judges as well. For the fun of it, we could also post leading entries and get feedback from people visiting botmag.com. Entrants would be asked to provide images, descriptions, and if appropriate, video clips and source code--enough information to enable us to make a fair and balanced evaluation.

I think some manufacturers may wish to participate, and prizes from them would fuel enthusiasm. We could also publish an article in print and on the web on the winning entries, and we could extend their subscriptions to Robot.

One idea that has surfaced in conversation would be the fastest line follower, with a predefined course that anyone could create with some electrical tape and a few square yards of highschool gym floor. That would level the playing field and nearly everyone has access to that. There are a lot of line trackers on the market and performance is all over the map. :-) Any thoughts anyone?

The main thing is to keep the barriers to entry low--the project would need to be affordable for typical hobbyists but also challenging enough such that a little ingenuity would make the difference. If it's a line tracker, we could have scratch-built and modified kit divisions.

So, forum visitors, give us your ideas, we hope to make a decision within several weeks if there is enough support and interest!

Tom Atwood, Editor-in-Chief

[I D Miller]

A line tracker is a great idea nice and simple i like the follow a string comment as well that could be interresting. I'm there where do i sign up

[dbc1218]

I think the line follower challenge is a great idea. The bigger idea of having a design challenge through the mag and over the internet is even better. There are no competitions for people who are not in a club, or don't have a club near them, or on a school team. This would be the first competion for everyone and I for one would be willing to compete.

[tatwood]

Today we are awarding our first robot photo and definition winners with subscription extensions and an honored place in our early magazine-sponsored-contest history; check out those items in this forum section, above this thread. But what additional challenges do we want to add?

This is your playground -- looking for suggestions from any and all--with a commitment to celebrate robotics exploration.

Tom
toma@botmag.com

[TheDuck]

Congratulations to the winners!

[ben-ja-men]

what about some sort of competition where a maze or whatever is set up at robot hq and competitors are given a protocol to opperate the robot with and are able to upload programs over the net to the robots to complete the competiton

[tatwood]

It's not out of the question that a designated platform could be used in competition and operated remotely by contestants. Plantraco had a robot tank on its website for a while that could be operated by visitors to the website.

For that matter, virtual robots might compete within the simulator in Microsoft Studio, or in similar programs if not in that particular one. But before we get into the heady world of teleoperated, internet-controlled or virtual robot contestants (should we consider setting up games in 2nd Life?), I think we should settle on some basic challenges that will get the compeitive juices flowing in the real world.

A roboticist friend some months back suggested that it would be cool to have a line-follower robot shootout. That made me think about a standardized track, and the logistics of obtaining the dozen or so line trackers out there. But why not start with a track and let anyone bring any vehicle to that track. Two RC cars with line tracking sensors were moving around a loopy track at RoboNexus in late 2005, and they were moving at a reasonable clip.

Anyone want to propose dimensions for a reasonable line "track" made out of black electrical tape on concrete or gym floors? :-) We might want to contact the people who made the cars exibited at RoboNexus and get a sense of course size and angles. I'll try to track them down and will be back with more. I believe these were 10th scale RC cars, stay tuned.

[wrighthobbies]

Hi All,

Reading through this thread, I see a lot of discussion about line following/line racing. I'm a member of Chibots and we do several line following contests including one that's designed for high speed bots.

We began with a simple track design years ago. As the club got better at designing line followers, we added in other types of line following contests including one for high-speed racing (against the clock). Some of us have achieved line followers that can achieve faster than 1M/sec speeds.

Here's a video of mine on a test track:

http://www.wrighthobbies.net/video/arty2-1.wmv

I'll post a picture and a detailed description of Arty in the bot pictures thread.

We have a lot of information available at Chibots web site - http://www.chibots.org.

[RoycePipkins]

Tom invited some of the folks in Chibots robot club to drop by this forum and talk about our experiences with running line following contests. We run four types of line following:

Basic Line Following
-Simple continuous course with gentle turns
Advanced Line Following
-Harsh course with zig-zags, broken lines and criss-crossing lines
Maze Solving
-a line maze laid out on a regular grid
Table Top Games
-a table with tap-marked end zones that a robot must travel between

One thing I'd like to note about running a line following contest is that there is a pretty wide range robot builder skill levels out there. First time builders will struggle to follow the line at crawl speed where as very advanced builders, such as the poster above, will zip around the course at speeds exceeding 1 m/s.

In order to encourage greater participation and enjoyment you may wish to separate the contest into experience categories in some way. It could be as simple a the builder declaring themselves to be a beginner, intermediate or advanced.

Particularly for the advanced builders, it will be important to define the race course pretty precisely. Floor material, lighting, minimum turn radius, etc. At our in-person contests it is quite common for the advanced builders to tweak control parameters on the bot on second and third chance runs to try to get an extra bit of speed out of it. For a remote contest they'll want to be able to replicate as much of the actual race course as is reasonable.

Just a few thoughts for you. Hopefully some of our other members will drop by with more suggestions for the contest.

Regards,
Royce Pipkins
Chibots President 2006
http://chibots.org

[hassenplug]

Anyone want to propose dimensions for a reasonable line "track" made out of black electrical tape on concrete or gym floors? :-) We might want to contact the people who made the cars exibited at RoboNexus and get a sense of course size and angles. I'll try to track them down and will be back with more. I believe these were 10th scale RC cars, stay tuned.

I think the dimensions of the car will depend on the dimensions of the track, and vise-versa.

Most of the line following competitions I've seen have turns of 6 inch radius. I'm sure it would be difficult for RC cars to follow such a line. They would need a larger curve. At the Chibots event just a few weeks ago, one of my robots would slide the front steering wheels as they went around the curve, at full speed. And that's only at 1 meter/second. I'm pretty sure RC cars could be a bit faster.

Steve

[jkjellman]

Hello, KJohn here (another Chibots member ),

While racing around a track following lines is great fun I actually prefer making the bot do a bit of thinking. The line maze we run at Chibots is a good example. This is a simple course laid out on 12 in square flooring tiles. White electrical tape is run down the horizontal and vertical centers of each tile. All the tiles are places on a 6 ft by 8 ft table to form the maze.

The tiles have basic line patterns, "L"s, "T"s, "+"s, etc. These are allow a grid like maze to be built. Our current maze contests has fixed starting and ending points on the perimeter of the maze and only a single shortest path from start to finish. There are no loops and the start and finish each have only one entry/exit line.

The thing I like about this is the is isn't just how good you are at the mechanics of building a bot (though this does help ;-) but also how well you can program you bot. For example, the maze solving robot I am working on right now makes it's first pass through the maze using a right hand garden rule. After this single trip through the maze it then calculates the shortest path to the finish so on it's next run it can take the direct route to the finish. I'd post a picture, but apparently you have to have a web site with your picture already on it to do so. :(

Anyway, my whole point here is let's make some of the contests involve finesse / guile and not just raw speed.

Take care,
KJohn

[jkjellman]

At our club (Chibots) we use 12 in square flooring tiles to build all our Basic Line Following (we have 3 types) and Line Maze contests. We use Black Armstrong Imperial Texture floor tiles #51910 which are available at Home Depot (and probably other home improvement centers) for less than 50 cents each. For the lines we use 3/4 in white electrical tape (available at most hardware stores or home improvement centers).

The reason for this is that it produces a very flat clean surface. The tiles do not transmit IR light and work well with visible light. They do have some grays specks on them, but the contrast between this and the shiny white vinyl tape is more than enough for all sensors we have run across (RIS,NXT, RIS, IR, etc).

The tiles can just be laid on any hard (spongy carpeting is not a good idea) surface to produce a fairly consistent background. If you make tiles with straight lines, crosses and 6 in radius turns, you can build just about any track given you have the space. As an added bonus if you turn the tiles with a single line sideways, you can create zones for a table top robot.

This would be a good candidate for the ROBOT line following competition as it is widely available (not sure about internationally :confused: ) and produces a very consistent surface. I'm not too sure how you would run a geographically diverse contest, possibly anyone could enter the contest by say, video taping their robot running the course and emailing in the submission. If speed is part of the selected contest, an analog clock should be in the video (does stop baddies who would tamper, but makes it a bit harder ;) ).

Anyway enough food for thought for now. Anyone else have any ideas on how to provide a consistently repeatable track for line following?

[wrighthobbies]

Most of the line following competitions I've seen have turns of 6 inch radius. I'm sure it would be difficult for RC cars to follow such a line. They would need a larger curve. At the Chibots event just a few weeks ago, one of my robots would slide the front steering wheels as they went around the curve, at full speed. And that's only at 1 meter/second. I'm pretty sure RC cars could be a bit faster.

Steve

One of my first bots was a modified RC truck. It was too fast and tended to roll over. Of course, I wanted it to go full speed all the time :)

I believe there are some RC cars designed to turn tighter corners. Wasn't there 1/28 or 1/32 scale RC cars at iHobby? I don't remember the details but they were turning some pretty tight corners. Those might be able to do 6" turns at greater than 1M/sec...

Another option is to use a pair of 45 degree turns like we do on the speed course.

[tatwood]

Just wanted to say that Robot magazine appreciates the input of the Chibots on all matters of Line Following competition.

Tom Atwood
Editor-in-chief, Robot
toma@botmag.com

[tatwood]

All who have followed the line following discussion on this thread may wish to view the four categories of line following competition that have been broken out in separate threads at top. This has been done at the suggestion of the Chibots robot club in Chicago. We invite your participation.

Tom Atwood
toma@botmag.com