FRC Robots Aren't Real Robots?

[Mikell Taylor]

Look, I'm a robotics engineer who works for robotics companies. If you get more than one robotics professional in a room, they'll all disagree on the definition of robot. The CEO of iRobot thinks a vending machine is a robot. I disagree. Drones are remote controlled, AUVs aren't, both are generally considered robots. FIRST robots definitely meet many commonly agreed upon definitions of robots. Some will disagree. Good for them. In my opinion, it's not worth arguing over.

[marshall]

Originally Posted by Mikell Taylor

Look, I'm a robotics engineer who works for robotics companies. If you get more than one robotics professional in a room, they'll all disagree on the definition of robot. The CEO of iRobot thinks a vending machine is a robot. I disagree. Drones are remote controlled, AUVs aren't, both are generally considered robots. FIRST robots definitely meet many commonly agreed upon definitions of robots. Some will disagree. Good for them. In my opinion, it's not worth arguing over.

Can confirm. Work in IT. Spend lots of time with manufacturing companies, hospitals, energy companies, and more. Engineers do not agree on anything.

Robots have sensors for input right? Well, what is a controller if not a set of sensors for input? It's a silly debate.

[SoftwareBug2.0]

Forgive me if you've heard me say this before, but I think FIRST needs to make a game where there's a huge penalty for not moving at all during autonomous. The number of teams that do nothing in autonomous every year is ridiculous.

As an example rule, let's say that driver control doesn't begin until you leave the starting zone of the field. You can either make your robot drive a few feet or hope that your teammates come to push you out of the zone.

There's no team where turning on motors for a few seconds is out of reach. If they can't it's because they've prioritized something else.

[mathking]

Originally Posted by SoftwareBug2.0

Forgive me if you've heard me say this before, but I think FIRST needs to make a game where there's a huge penalty for not moving at all during autonomous. The number of teams that do nothing in autonomous every year is ridiculous.

As an example rule, let's say that driver control doesn't begin until you leave the starting zone of the field. You can either make your robot drive a few feet or hope that your teammates come to push you out of the zone.

There's no team where turning on motors for a few seconds is out of reach. If they can't it's because they've prioritized something else.

I get the frustration that many of us feel when many robots don't move during autonomous. This rule at first glance sounds good, but there are some practical problems to think about. The first is that you would need to define the game in the right way to make this practical. In many games there is a substantial penalty when your robot interferes with another robot's autonomous mode. In others there are potentially really bad mistakes that can happen in autonomous that should be avoided. Some teams opt to do nothing not because they can't but because they don't want a mistake to cost their alliance. We have had sensors that prevented a collision during autonomous, making our robot not move very far.

Another consideration is that this rule will definitely lead to more matches (potentially many more) where robots sit and do nothing. There is nothing more frustrating in FRC competition, for any team but particularly for new teams, than having a dead robot.

Yet another consideration would be that sometimes the field communication system can mess up a team's autonomous mode. We had at least one match this year where another team not being able to connect to the field led to a reset of the communications. Our autonomous mode had been selected, but when the field reset this choice was lost. The robot did nothing, even though it was consistently scoring. We ended up losing that match 118-119. That was frustrating but nowhere near as frustrating as it would have been if as a result of a field reset we had not been able to move at all for the entire match.

I think if you want to increase the number of robots that do something in autonomous the best solution is to provide a sufficient incentive to get teams to do something. In general positive incentives tend to be more effective in game theoretic / behavioral economic motivation anyway. (Humans tend to underestimate the chance of events triggering negative consequences and overestimate the chance of events triggering positive consequences.)

[Caleb Sykes]

Originally Posted by SoftwareBug2.0

Forgive me if you've heard me say this before, but I think FIRST needs to make a game where there's a huge penalty for not moving at all during autonomous. The number of teams that do nothing in autonomous every year is ridiculous.

As an example rule, let's say that driver control doesn't begin until you leave the starting zone of the field. You can either make your robot drive a few feet or hope that your teammates come to push you out of the zone.

There's no team where turning on motors for a few seconds is out of reach. If they can't it's because they've prioritized something else.

I think the GDC tried to implement something like this in 2015, and the result was that even fewer teams bothered with autonomous at all.

This year struck a good balance in my opinion. 2 points just for running motors a few seconds, and 3 more bonus points if you could just make it over a defense. I wouldn't mind seeing some form of disincentive for smashing into the opposing alliance wall at full speed though.

[EricH]

Originally Posted by Caleb Sykes

I think the GDC tried to implement something like this in 2015, and the result was that even fewer teams bothered with autonomous at all.

This year struck a good balance in my opinion. 2 points just for running motors a few seconds, and 3 more bonus points if you could just make it over a defense. I wouldn't mind seeing some form of disincentive for smashing into the opposing alliance wall at full speed though.

My take:

If you don't leave zone X (in the right direction) by the end of auto, -2 points/robot AND temporary disadvantage. If you do leave, +5 points/robot. If you leave in the wrong direction, then that can be discussed later--game-dependent.

The temporary disadvantage would be that for 10 seconds, you are stuck in auto unless you leave the zone (by any means). At 10 seconds or zone exit, you're put into teleop.


That being said, points for running your motors is good, with more points for doing more being better. But I suspect that a mild point penalty for not moving in auto--and a piece of code, from FIRST, that is set up for "move so far"--would provide a large incentive to teams to make sure everybody moves in auto.

[SoftwareBug2.0]

Originally Posted by EricH

My take:

If you don't leave zone X (in the right direction) by the end of auto, -2 points/robot AND temporary disadvantage. If you do leave, +5 points/robot. If you leave in the wrong direction, then that can be discussed later--game-dependent.

The temporary disadvantage would be that for 10 seconds, you are stuck in auto unless you leave the zone (by any means). At 10 seconds or zone exit, you're put into teleop.


That being said, points for running your motors is good, with more points for doing more being better. But I suspect that a mild point penalty for not moving in auto--and a piece of code, from FIRST, that is set up for "move so far"--would provide a large incentive to teams to make sure everybody moves in auto.

I like the way you're thinking. There's some room for creativity in the autonomous mode rules. It's rarely something other than gain a few points for completing easy objectives, more points for hard objectives, and a big penalty if you interfere with the other alliance.

As for the specifics, I like the idea of losing some points, and not just gaining them. In games with the scoring levels we've seen recently, 7 points probably isn't enough though. 2014 had 5 points for driving over a line and a lot of teams didn't do that. My first thought had been to make moving in auto worth like 50 points, but then that might make games where one partner was missing unwinnable from the get go. If instead there was a penalty for robots left in an area the alliance isn't totally hosed.

I also like the idea of a temporary disadvantage. This is something FIRST has done before. In the 2004 game if a certain action didn't happen in auto then gamepieces were held back for a certain number of seconds.

Also, as mathking brings up, there may be some teams that don't do anything for fear of messing things up. One way to reduce this is to eliminate penalties for interfering with the other teams during autonomous.

[Type]

Originally Posted by SoftwareBug2.0

Forgive me if you've heard me say this before, but I think FIRST needs to make a game where there's a huge penalty for not moving at all during autonomous. The number of teams that do nothing in autonomous every year is ridiculous.

As an example rule, let's say that driver control doesn't begin until you leave the starting zone of the field. You can either make your robot drive a few feet or hope that your teammates come to push you out of the zone.

There's no team where turning on motors for a few seconds is out of reach. If they can't it's because they've prioritized something else.

I think it's a good idea but at the same time I disagree. There have been matches where they start the match without us, even when we were banging on the glass, so we didn't get a chance to do auto since we didn't have time to select one, I'm notrying even sure if we had codecided loaded from the Rio, like I don't know if it was reading it. If they waited until everybody was ready like they are suppose to, it may work, but things can mess us and the auto on a robot may not start.

[SoftwareBug2.0]

Originally Posted by Type

I think it's a good idea but at the same time I disagree. There have been matches where they start the match without us, even when we were banging on the glass, so we didn't get a chance to do auto since we didn't have time to select one, I'm notrying even sure if we had codecided loaded from the Rio, like I don't know if it was reading it. If they waited until everybody was ready like they are suppose to, it may work, but things can mess us and the auto on a robot may not start.

I'm not sure that I totally understand what you mean, but couldn't you solve your problem by having a default that's not "do nothing"?

[Type]

Originally Posted by SoftwareBug2.0

I'm not sure that I totally understand what you mean, but couldn't you solve your problem by having a default that's not "do nothing"?

It was awhile ago so I don't remember it all but our default auto didn't work either. I wanna say we weren't connected to the field but it seems like they would have noticed the light. I was just trying to say in a world where everything worked as its supposed to, it would be a good idea but there are times where the field messes up.

[mathking]

Originally Posted by Type

It was awhile ago so I don't remember it all but our default auto didn't work either. I wanna say we weren't connected to the field but it seems like they would have noticed the light. I was just trying to say in a world where everything worked as its supposed to, it would be a good idea but there are times where the field messes up.

The problem we, and others, have had is that if you use the dashboard to select an autonomous mode, and the field has to be reset in order to get another robot to connect, you can lose your dashboard selection of autonomous mode. This happened to us at least once this year where when the reset happened we lost the selection and the drive team was unable to get the attention of the field crew to wait.

There are a number of different ways the the FMS to robot interaction can cause problems. (This is not a complaint about the FMS, so please let's not start that discussion in this thread.) Any sort of complicated communication system like that is going to have problems from time to time. I think that punishing a robot for not moving by making it not move for an even longer time is not the way to go. That would make one of the most frustrating situations in FRC, a robot that is unable to move, far more common. It would also disproportionately disadvantage inexperienced teams. Adding incentives to make it worthwhile to move is great. And avoiding game rules, such as we had in 2015, which encourage some teams to tell others not to move as a standard plan is also a good idea. But please let's try to not to increase the number of times robots can't move.