What Does Everybody Think About Using Sensors In The Competition For This Year?

[Hailfire]

What do you all think about using sensors, programming them, etc. for this years competition?

[Jeff Waegelin]

I think it's a pain, but it does definitely make it interesting. It'll sure give me more work to do...

[Hailfire]

Yeah, especially programming it to move by itself for 15 seconds.:p

[Noah]

I think that with 15 seconds of automation, you have a chance to do real damage if your programmed well and the enemy is not. Part of a good automatic program, however, comes from sensor input, and should definitly not be overlooked.

[Mongoose]

I think it's pretty cool, because last year I don't think we (our team) focused much on the software side, if at all. Plus I'm interested in AI and that sort of stuff. The part that I fear is wiring everything into the analog/digital inputs.

[Alexander McGee]

well, i think its a cool idea, but, its also a new thing this year. i dont think that all teams are going to use it, or use it well. well just have to wait and see.

[Joe Matt]

It'll add another challagne to the game that other robotics competitons don't have. Also, I expect loads of people to use the reflective tape on the bins.

[Norm M.]

I suspect there will be a lot of programs that have the robot take a look to see if there are any significant stacks (3 or more). If yes, give them a shove. If not, make a u turn and bulldoze the wall of containers onto your side of the field.

Certainly not a trivial task. This is gonna be fun!

[Ryan Meador]

Norm, that's a lot more complicated than you make it sound ;) But I love the spirit. I've been pretty much floating around ever since the kickoff... I've been waiting for exactly this move for years. We're all supposed to get 15 minutes of fame, but I guess us programmers only rate 15 seconds. Anywho, my real point is this: sensors are king, and how you use the first 15 seconds I suspect will have a huge impact on who's the winner...

[Chris_C]

Indeed. I think FIRST has taken an interesting risk this year with giving us programmers the first 15 seconds of the game. In some of the posts on the general forum, people seem to be agreeing that a well programmed robot can basically decide the game in that time period. I also think that the light sensors are going to be a bit trickier to work with than everyone is giving them credit for. I wouldn't call any of the robots that my team has built over the last few years precision machines, and you are going to need quite a bit of precision to get your robot to move around based on light sensor readings and perform accurate movements while completely automated. Despite the challenges though, I am absolutely thrilled that the people at FIRST are giving us this opportunity. As my team's main programmer, this change is going to make the season a whole lot more exciting for me.

[Ryan Meador]

Amen. I, too, am my team's main programmer. There was a time when I was the only programmer, but now that I'm a senior and we have others, I'm the programming coach. Good deal :) At least I'll go out with a bang. I think most people are fretting too much about this. I used the optical sensors last year. They're pretty nifty, and not hard to work with at all, as far as hardware goes. And the software is only slightly more complex. I'm not a big fan of precision movements or the like (just ask my teammates... they'll assure you my style is more to throw a bunch of stuff together and see what life crawls out of the pile). I believe less planning and more correcting is the way to go. That is, dash off in a likely direction and keep correcting your direction as you go, homing in on the target. Think like a cruise missile, not an artillery shell ;) Good luck all... this is going to be a great year.

[Skabana159]

My team is already coming up with ideas to baffle me and my programming team. They think we should zip onto the ramp and push boxes over within 5 seconds, all the while looking for stacks, and if one is over three, go for it after we fininsh pushing the middle ones over. This is going to be fun. We might have to resort to "dead reckoning" if following the line isn't fast enough.

[rwaliany]

They're pretty basic once you have them hooked up. If you're trying to create a line follower simply attach a light sensor to one of the digital inputs. Check whether the number one corresponds to the white line or the other material next to the white line "Debug ? rc_sw(switch)". If they are the same then adjust the circle knob thing on the back of the light sensor till one is invoking one and one is invoking zero. Once thats done, then use an if statement and make the robot move left and forward if it sees the whiteline and right and forward if it sees the other surface.

If you want to be more advanced you can use two light sensors, which will allow you to move faster, and if one of them sees the white line turn the opposite direction. If they are both giving the no-white-line signal then go forward.

[Raven_Writer]

Originally posted by rwaliany
They're pretty basic once you have them hooked up. If you're trying to create a line follower simply attach a light sensor to one of the digital inputs. Check whether the number one corresponds to the white line or the other material next to the white line "Debug ? rc_sw(switch)". If they are the same then adjust the circle knob thing on the back of the light sensor till one is invoking one and one is invoking zero. Once thats done, then use an if statement and make the robot move left and forward if it sees the whiteline and right and forward if it sees the other surface.

If you want to be more advanced you can use two light sensors, which will allow you to move faster, and if one of them sees the white line turn the opposite direction. If they are both giving the no-white-line signal then go forward.
Following the line isn't exactly the best thing to do when you think about it. The robot will move slower because it has to constantly follow the path, which is what you don't want.

[rwaliany]

I'm just showing a simple algorithm to do it. Plus, most teams I'm thinking will create a couple of different programs and switch it based on their alliance partner.

if you use two light sensors and have a good distance away from the line, and go straight until a hit, then turn until it hits on the right one. then go straight and perform same algorithm, it's pretty efficient. You'd do 5 turns at max. What would you choose accuracy or speed.

Doing it time based is still 3 turns. Well you could always use the radius of the circle and make a triangle using delta_t to brace time. But I would still use a light sensor to check if your on track every second or so.

[Noah]

Following the line isn't exactly the best thing to do when you think about it. The robot will move slower because it has to constantly follow the path, which is what you don't want.

I agree. While a line tracker is a good thought, I think it would be faster to go dead reckoning around and up the ramp. And besides, all the line is good for is getting you on to the ramp anyways! beyond that, if all you have is a line tracker, your a dead reckoning bot!

[Raven_Writer]

And everyone knows the majority of the teams are gonna make the robot go up the ramp first, so it's a matter of speed and strength.

On a side note though, it'd be sweet if we could do flippin' this year, does anyone know why they banned it since they already said "This game is not ment to be fair."?

[rwaliany]

Dead Reckoning...faster...so unreliable. I hope the materials are the same on the field as the ones you practice on.

I don't know, our team is going to have 2 bit switches for 4 automaton programs. The best way is to test it out, seeing if accuracy is more important to your team than speed.

What if your alliance team member has a dead reckoning program too. Do you guys run into each other and slow each other down, while the other team comes over the bridge. Better yet, you hit, your now off target and going the wrong way.

I'd say using the light sensors for the TURNING, then once your on the top use a yaw rate sensor to detect going up the ramp, to stop the light sensor turning.

Right now based on our opponents/allies we're going to have
One line follower,
One dead reckoning,
One seek and destroy,
and One salvager or something

[Raven_Writer]

Accuracy is a better power to have than speed, but it's true. It's gonna be like a demolition derby or something.

[Ryan Meador]

I believe an array of 5 sensors (probably not the normal optical ones, but "custom circuits") would make it very fast, if arrayed two on the outside, and three clustered together at the width of the tape between the other two. You can zip around and use the three for precise driving as much as possible, but if you lose the line at high speed, you've got the other set to follow the line more loosely. I haven't tested this yet, but I'm going to as soon as I get at last year's robot ;)

[HolyMasamune]

That's why communication between teams is very important as well as multiple codes.

[Raven_Writer]

at least 3 (2 for automation (or however it's spelt), and 1 for the 2 min. part)