![]() |
Re: Drive Team Configuration
I am not going to judge on what one team does or does not do, but I will provide an example.
In 2009, we were grouped with teams 1051 and 1746 in the eliminations. During the semi-finals. We won the first match, and lost the second match. In the third match, I noticed the referee throw up a penalty flag in the middle of the match, but was surprised to see that no penalty was assessed in the score. It wasn't until about 10 seconds later that we found out that our alliance had been disqualified because the coach on one of our teams touched the controls. The team in question ran three separate drive teams, and the coach on the drive team in during the match in question happened to be the driver in another drive team set, and when he noticed something wrong, his instincts took over and he reached in to adjust a potentiometer. Now we don't blame this team at all, they performed extremely well and we were grateful for everything they did, but one can't help but think of what could've been. Of course, I'm not against substituting a bad driver or a driver who seems to have a lot going on in his or her mind, but I do like consistency, especially when it comes to developing skills and general know hows of a robot and a game. Anyways, enjoy the break, and here's to a good 2011 for everyone!!! - Sunny |
Re: Drive Team Configuration
Quote:
Your failing to see the point. Adam suggested that you use your programming skills to help the drivers and the TEAM . With your abilities you could: -Create a competitive autonomous routine -Allow better control of the robot for drivers -Implement more sensors that would provide valuable feedback -Design a dashboard for you team. -Automate tasks that would be hard for drivers (such as re cockk the kicker after kicking the 2010 soccer ball and having the kicker go to a desired position). There are so many different and valuable things you can do to put your team in the best position to win. I believe there was a thread talking about how missions to the moon/mars are not fully autonomous and require some sort of human interaction. A fully autonomous crappy robot isn't really that amazing, a well built winning robot is hands down much more amazing. Its very foolish for you to ignore the advice of very talented and respectable CD members. To be honest, my team has worked all year long. Even in the off season. If my team was in the top 8 and we lost a spot cause our alliance partner used us as a "guinea pig", I would be very very upset and so would many of our parents/mentors. Taking everyones advice here on CD as a challenge makes you look quiet stupid, it doesn't hurt for you to stop an say "Man I was wrong" and take a new direction. Adam said it best: "In engineering, it's important to use the appropriate tools and not more resources than are required. Building a completely autonomous robot for a teleoperated competition may be cool, but it's bad engineering. A robot with many automated and autonomous portions to make it easier on the drivers (while maintaining driver control), is good engineering." Also take some time and think about the posts, don't reply back right away cause you HAVE to leave a reply. -RC |
Re: Drive Team Configuration
Quote:
By playing the game, I don't mean, followed a ball until it's picked up, then turned towards the goal and shot. I mean intelligently finds the closest or easiest game piece to grab while avoiding obstacles, knows not to shoat with an opponent in the way, and finally but precisely finds the goal with the game piece. Anything less than that really isn't playing the game. Yes, that probably wouldn't be the greatest use of human resources, to use a talented programmer to recreate the most basic human driver, but that is award winning (Innovation in Control...) and a valuable experience for the programmer in question. That said, it would probably hurt the team to a certain extent, and I can't imagine a potential drive team would be happy to see their places taken by a less competent AI. So David, if you really are going to try this, which I believe you will, if your AI is worse than a human driver, don't expect it to be used. However, if it is actually better, then that is a significant accomplishment. |
Re: Drive Team Configuration
Quote:
It's your duty in FRC to your teammates and alliance partners to put the most competitive robot possible on the field. A fully autonomous robot (assuming you do it, as literally hundreds of posts about the difficulties of such a system have again gone completely over your head) will in no way be better than a teleoperated robot with only 6 weeks of work. Though honestly, you're lacking something far more important to your alliance partners than a good robot: good listening skills. Though, hey, everyone else in the world is your guinea pig anyway, you won't need to coordinate with them. |
Re: Drive Team Configuration
David. Talk to 33 or 111. They have some of the most automated robots in the competition every year. They could probably make a serious run at a fully-autonomous robot if they wanted to. They don't. There is a reason for this!
33 has automation built into a lot of functions. A filter they wrote years ago went onto a sponsor's vehicle a year or two later with little modification. They were one of the first teams to do shift-on-the-fly--with a 4-speed gearbox that shifted smoothly. They write the automation code so that the operator can tell the robot what device X needs to be doing, and the robot just does it with no further input, if it can be done safely. Read the 2007 Behind the Design book. 111, same thing. They've built some complex robots, and the code to match. They did a 4-5 joint arm a few years back. (See the same book.) If they didn't have some form of automation, they'd be crazy. 1024, back in 2008, ran a full avoidance program (until the sensors smoked and flamed). They could avoid just about anything autonomously. To run a fully-autonomous robot and run it well, you need to combine the avoidance code of 1024's 2008 robot with the automation of 33 and 111, along with the decision-making of a human and the drive code of your robot design. If you can't do that decision-making, settle for automation. Make it so that your driver or operator says, I need X at point Y, and the robot does it quickly and smoothly. Very few robots could run fully autonomously these days. The last game that was practical was 2003: Drive up the ramp and lock down. 111 was one of the World Champions. This was also the first time automode came into play. 2002 would be even easier--drive straight, grab goal, drive straight, park. Imagine if 71 had run autonomously that year: Drivers hit "go" and watch the robot win the match every time (as opposed to drivers drive the robot out, shift drivetrains, and win the match). Listen to JVN. He's been on three good teams and learned from many more. If he says that he does not want to be paired with you, without even hearing about your robot, you might want to listen to why. On the same note, I'd want to pair with 33, 111, 71, or 148 any day, no matter the game and no matter their robot. There's a very good reason why, or several, and some of them are above. |
Re: Drive Team Configuration
Quote:
Going back to the FIRST acronym, For INSPIRATION and Recognition of SCIENCE and Technology, a working, functioning autonomous robot is hands-down the best way, IMO, to get people to be inspired by the competition. Interfacing with a human operator is a fine thing and all, but forcing a robot to run on it's own, and run well, is pure science and pure inspiration. People will respect that better than any winning robot that can do some things neatly and win, because FIRST and the FRC isn't about winning. If he can convince his team that this course would suit them well, and they feel like it's in their most inspirational interests to do so, they should. It's why they are building a robot, after all. |
Re: Drive Team Configuration
Quote:
Another note, how many FRC robots have successfully completed their autonomous program repeated in the exact same way every match? Along the same idea, how many robots have done the exact same things in each and every match? FLL robots compete on tables with no other robot interaction and most of them cannot pull off 400 points consistently every round all the time. I am not doubting your programming skills, it is A match strategy, but it isn't the ONLY of most EFFECTIVE strategy. Consider the amount of time it would take to program/test/debug all of that code in 2 minute matches. I would rather take all that time to train drivers how to play the game and react to different strategies. 1519 has done the usual driver for the chassis and moving the robot, operator for the manipulators, coach for direction, and human player to do the human player things. In 2009 we gave the control of picking up balls to our driver instead of our operator too ease communication between the two. Merry Christmas!! ***http://www.mechanicalmayhem.org/teamvideos.asp |
Re: Drive Team Configuration
Thanks to all of you that responded to my post, but now lets get back to the topic:
I suggest that you have the maximum number of persons allowed at the box at all times. You never know, what if the driver cramps up his forearm an cannot properly drive? Somebody has to come in and replace him. Ok far-fetched example, but I am exemplifying something. Always come prepared. You would never go to the regional without spare parts, tools and a computer to program the robot with. Why step onto the ring without a backup? Ben Franklin once said that failing to prepare is preparing to fail. You never know when your driver passes out during the match due to dehydration or his arms cramp up. Always have depth in your rosters. |
Re: Drive Team Configuration
Quote:
It is about winning and what that can mean. Jane |
Re: Drive Team Configuration
Quote:
Quote:
That's even before considering the increasingly obvious fact that the programmer in question (and his entire team, for that matter) has more than a few other things to work on before getting to the point where a fully autonomous robot is anything more than a fantasy. |
Re: Drive Team Configuration
816 has always followed the following configuration for as long as I've been involved and from what I've heard since our inception.
Driver - Responsible for the movement of the robot and getting the robot to where it needs to be. Usually we've tried to keep the driver's controls to a minimum so there is less to worry about, but we have played on giving the driver some control over other machine functions (Ramps in 2007, Dumper override in 2009 which was never used.) Operator - Responsible for robot manipulators. Pretty straight forward, anything that's not in the drive-train is usually controlled my the operator. Human Player - Varies with the game, and usually is the person most likely to change from year to year depending on the role. Usually this person is relatively athletic should the game call for it, otherwise this role is assigned to one of the more dedicated members that wasn't able to cut it as Operator or Driver. Coach - This is the role I've somehow fallen into since my graduation from HS, and I take responsibility for pre-match strategy, Analyzing Scouting Info (This varies from year to year, depending on what human resources are available) and watching the match while it is in progress. I try to keep my driveteam as well informed as possible both before and during the match so that they have less to worry about. We also have a sort of an Assistant Coach on 816, Zach. He was our Coach in 2009 and a really close friend of mine so he usually travels with the team to help with strategy, scouting, keeping me sane, match analysis, comic relief and a handful of other things. If you don't have a Zach on your team, I think you should find one. I place a large emphasis on drive team cohesiveness (I guess that's the appropriate term) and how well the personalities mesh with each other. I've worked with a lot of different drive team configurations in my 5-seasons with FIRST and I've found that this is often over looked. In many years we'd often pick the person that was best for a single job but often conflicting personalities would hinder the drive team from reaching it's full potential. Just something to think about. We also like to keep at least one or two back ups for each position, but we never cycle through drive teams. (anymore) Consistency is often one of the most important things when Alliance Selections come around, so you should keep this in mind. Nothing's worse that picking a team and having them switch drivers or coaches during a match - it really messes up the flow of things. Quote:
I really might have to borrow this one. |
Re: Drive Team Configuration
In 2010 for 1178 we had:
Driver: Drove the robot, could activate the kicker, or hold down a button to deactivate it Weapons Operator: (Team Name, makes it sound really cool) Could kick, deactivate kicker, change kick setting and activate the winch on the robot. Coach: Talked strategy with the other teams before the match, keep track of the score and penalties, help the human players communicate (if the person on the trident wasn't paying attention) and sometimes would tell us what our alliance partners would do Human player: Pulled the balls out of the corral and handed them to the person with the trident We kept the same four people for the whole St. Louis Regional, and aside from robot problems, it worked fairly well. I cannot tell you about 2009 but in 2008 the driver was given more control of other functions as the Weapons Operator had a complex task. |
Re: Drive Team Configuration
Quote:
|
Re: Drive Team Configuration
Quote:
|
Re: Drive Team Configuration
Quote:
Hey David, I believe I can safely say that I know where you're coming from with your confidence in achieving a fully autonomous robot, and like some of the CD programmers here, I wish you the best of luck in achieving that. I've read several of your posts in this previous year and although I haven't seen any of your works personally, you've given me the impression of a very talented person, aside of being a talented programmer. I also assume that you are more experienced in this field, more than in mechanics. As such a talented and experienced programmer, I'm quite sure you realize that by looking at previous years games and the diversity in the challenges and tasks in most of them, especially from every following year, you can see that it is very difficult to predetermine the challenges your TEAM'S robot will have to go up against. You may predict and prepare, and sometimes that works, but as for automation that is nearly impossible. There are a lot of FRC rules, field parameters and game element parameters and robot parameters and much more that you must take into consideration when forming up your TEAM'S strategy of the autonomous robot. But wait! Now you might know the rules, but you need to define GAME STRATEGY of your TEAM'S robot. Now that should put in some more considerations. Can you define all of the possibilities of different types of strategy-based tactics? Tons! Assuming you can define the solutions and implement them into the robot while in the competitions is very unlikely. Speaking of parameters, somewhere deep in that huge paragraph i just worte (:p ), Can you really trust your TEAM'S robot's sensors and their accuracy? What if changes are made to the robot's physical model? Out of experience (as myself and as observing the work of another talented programmer I personally know), you may have the whole game plan for programming and operating the fully autonomous robot, but there's a huge consideration you need to take in mind. What if the rest of your TEAM cannot build the robot your TEAM's strategy is based on? What if certain MECHANIC components or ideas just don't work? A lot of time (and that is part of merely 6 WEEKS), and that also means sensor and manipulator calibration time, testing time and correction time (which includes going over the last two again), will be lost if that happens, and, sadly, as the young engineers and fragile humans we are, it happens to us a lot; And as a programmer, it's a real shame to think that you might have had the "ultimate code to rule them all!"...when even in practice you couldn't bring it out in the final product, the TEAM's robot. I have already experienced in the past 4 years the disappointment when you realize you've written a great code for your robot's mechanical component, but it isn't built until the very last days, and sometimes is just scratched off the final product, and all that code has gone for the season. Your autonomous robot relies HEAVILY on the REST OF YOUR TEAM to build the other parts of your robot successfully, that is the electrical and mechanical parts. If they can't achieve even the minimum required out of an FRC robot, your code is just virtual for the season, which is the most important period - not the off-season. Virtual is beautiful and awesome, but reality is much more awesome, more practical and more accepted within the ranks of engineers and scientists in the field future science and engineering. That was more of a personal statement for you David, and now to the point: Unless you have 100%, or even 95% assurance you can achieve your goal with all it's prior requirements to achieve it, then go for it. But if you cannot assure it, it's best to lower your goals until you reach that 95-100% area of insurance of success. Like others said, helping out your drivers, who will need to manually drive the robot in case of a "fully-autonomous emergency", by giving them easier and smoother controls, faster response, simpler manual control methods and more, which are much more achievable and mostly likely to guarantee your TEAM success when driving the robot. Like I've said, you're probably a very talented programmer and can probably think of very creative and useful ways to help out your TEAM in driving the robot, and hey!, maybe even when building it, say for mechanical tests! ...Though on the other hand you do have 8 programmers on your team, which is already too much... :S Non-the-less, best of luck to you and your team in this year's season, and I hope to hear great stories from your side. :) |
| All times are GMT -5. The time now is 23:46. |
Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi