The radius of the shooter arc is 13.5", and the height between top tracks and bottom of the shooter is 1.5".
Why did you all move away from the Versawheels to the AM HiGrips?
We drove the robot hard in the last couple weeks of build and wore almost all the tread off our first pair of Versawheels. With no practice bot our robot saw a good amount of driving practice before ship day. Our drive was designed in such a way that changing wheels was not a trivial task. We thought we could get a regional or two out of new Versa’s, but really didn’t want to change wheels again at Championship.
I also remember hearing that when you were practicing you tore the tread right off of one of your wheels. Nick Lawrence was playing defense I believe, and, well… He wasn’t going easy on them.
Austin and his partner were both great drivers!
During the days before we had rocker (I know! :ahh:) we were doing some driver practice against last year’s robot driven by Nick Lawrence and Kaj Anatharajah and we tore a Toonie-sized (Canadian $2 coin) hole in the tread on one of the wheels. With this and a few other issues, we took the robot back to the shop, added rocker, and changed out the wheels to a new set of VersaWheels. Even with the rocker, however, we still wore down the wheels too quickly to continue using them when they took so long to change. This in mind, the first thing we did at BAE was change to the AM HiGrips. We later ran in to Mr. Copioli at Waterloo who gave us a much needed lo-down on the VersaWheels (as well as a priceless reaction upon seeing our damaged wheel). He basically told us that the Versa’s weren’t meant to be the do-all, end-all wheel like many of the AM wheels, but more of an easily-replaceable and cost-effective option with high performance (which - at $5 each - they are). Our drivetrain style just didn’t lend its self easily to the way they were meant to be used. If we ever go cantilevered, the Versa’s will be near the top of our list for wheel choice.
Thanks for this, but a lot of credit is due to both Mr. Lim as our drive coach and (especially) the programmers for how well we were able to drive. I know that the programmers made my job hundreds of times easier than it could have been. In fact, they made it so easy that some have even dared to say “anyone can operate”.
Could you elaborate on what the programmers did? Outside of modifying the joystick output line into a curve, I’ve never liked modifying tank-drive controls because it typically limited driver capabilities.
Oh, and while I’m posting, I’d love to hear how you guys had such a quick firing rate. Was it the double mini-CIM combo, or was it something else, like an RPM sweet-spot on the wheel?
Outstanding job, by the way, and congratulations on your Einstein win!
I’m not the best person to answer this but I do know a little bit about how the robot was coded from different points in the season. First, we use Logitech USB gaming controllers (PS3 style) for both the operator and driver. The control system we use is known as “Kaj Drive” lovingly named after mentor Kaj Anatharajah. Generally it is referred to “Halo Drive” by other teams but it consists of the left joystick controlling forward/back motion and the right joystick controlling left/right rotation. I used to prefer traditional tank drive as it lets you control only half of the drive at once allowing you to get out of sticky situations easier, but Kaj is so much more intuitive that I had to convert. We also cube the drive functions so that we have finer control when nudging the joysticks.
The double mini-CIM combo did help our fire-rate, but we were still shooting pretty rapidly before we re-did the shooter for champs. What gave us an edge on shooting was our very clever but relatively simple shooter code. I don’t have a code snipit, but I can explain it. First, we were running the shooter for the whole match, at just under our target shooting RPM for both of our shooter’s positions. Low position (full court and loading), and high position (back-centre of the pyramid and radius with the goal). When I pressed and held the trigger, it would give a full 12+V to the shooter motors, and then trigger the feeding piston when our optical encoder got the target RPM. This means that we could shoot almost as fast as our wheel could recover. Again, the extra mini-CIM helped our shooting a noticeable amount, but was not the sole contributor to our shooting success. You can really see the difference in the two videos below:
BAE Finals 1: http://www.youtube.com/watch?v=lvjdNey2IlM&t=0m32s
Einstein Finals 3: http://www.youtube.com/watch?v=MQ0VNzIvHx0&t=2m15s
Thank you and thank you! It was good to see you guys recognized with a Chairman’s award after so long as well!
Matthew gave an excellent overview above. Drivetrain driveability is critical, and some time goes into simply getting the robot to “feel right” per the driver’s feedback. Optimal transfer functions change from year to year. Kaj drive with a cubic transfer function worked for us this year.
Shooting precision, accuracy and recovery time was another big focus. We needed it to be as simple as holding down the shoot button, and the robot would deliver all 4 discs as quickly and precisely as possible. We tried various flavours of PID, and ultimately settled on some full-power + open loop trickery. The shooter was held to 100 RPMs lower than our desired shooting speed. When the operator pressed the shoot button, we applied full power to the shooter, and the feeder fed a frisbee precisely when the desired speed was being surpassed.
A fair number of checks had to be coded around this to handle mis-feeds, to account for how fast our feeder could actuate, and to otherwise make this “foolproof.”
Furthermore, there were several “active trimming” functions that allowed the operator to “live tune” the shooter throughout the match, to immediately adjust for any missed shots.
Lastly, little details like mapping the most commonly used functions to the trigger buttons allows you to actuate them while simultaneously driving/trimming. Also, having handheld gamepads allows you to walk around inside the driver station while driving. It helps to be able to take a few steps to look around a robot that’s blocking your view. Few teams realize that your drivers (and coach!) are free to roam the entire alliance station (not just your team’s 1/3), if they need to.
Teams in the past have asked to see our code. We’ve been using Google Code for the past few years for version control, which means it is publicly available. If you search for “robotics610” there, you can see the nitty gritty details that our students put into our code this year. I never say anything we do is the “flashiest” or has huge “wow factor,” but I am extremely proud of the time and effort they put in to doing the simple things as best as they possibly can.