2010 5th Gear FRC Match Simulator

[Bjenks548]

Originally Posted by Rangel(kf7fdb)

Does anyone know how to maximize the window or somehow make it full screen even though the maximize button is disabled?

Yea, theres a full screen option under graffic settings

[gblake]

Folks,

Not too long from now I think I'll be in the market for a beta tester or two for some Java 5th Gear work I'm doing. Can I get a couple of volunteers?

I think I will give you some pre-compiled class files and let you tell me how easy (or horrible) it is to complile them into a application you write to use them.

Part of this will include me supplying good enough written descriptions for what the classes do.

Before any student volunteers, they must have all of their school work under control (You already have a guaranteed A in all of your classes or you have already taken all of your final exams).

These software classes are less important than school classes; and won't be ready for a couple weeks or so anyway.

Blake

[BrendanB]

Originally Posted by gblake

Folks,

Not too long from now I think I'll be in the market for a beta tester or two for some Java 5th Gear work I'm doing. Can I get a couple of volunteers?

I think I will give you some pre-compiled class files and let you tell me how easy (or horrible) it is to complile them into a application you write to use them.

Part of this will include me supplying good enough written descriptions for what the classes do.

Before any student volunteers, they must have all of their school work under control (You already have a guaranteed A in all of your classes or you have already taken all of your final exams).

These software classes are less important than school classes; and won't be ready for a couple weeks or so anyway.

Blake

Just for clarification, is that testing software, programming, or a combination of both? I am a little confused.

New score of 29. This simulator is so much fun, thank you so much for making this!

[gblake]

Originally Posted by BrendanB

Just for clarification, is that testing software, programming, or a combination of both? I am a little confused.

New score of 29. This simulator is so much fun, thank you so much for making this!

First - You are welcome - Mentors never tire of hearing that.

My big picture goal is making the simulator more of a learning/mentoring tool than it is already. The software I want to release will be a lure meant to entice students to do some fun coding on their own.

What I will supply is a a set of classes that can communicate with a 5th Gear server. They will also expose a) (in a list of objects) all the data a 5th Gear client gets about the states of the simulated game objects, the score, and the time left in a match; and b) an API you can use to feed robot-control commands into the server.

Right now I intend for my code (in a single computer) to be able to replace from one to six clients .

I will have some cheesy 2D graphics built in for simple visualization of where the game pieces and robots are on the field.

On your end of things, I am hoping for three sorts of beta testing 1) Can you pull my classes into a main routine you write and successfully interact with them (display/analyze a robot/ball's location, movement, etc.), 2) can you issue orders to a robot (move, turn, grab, kick), and 3) does my simple GUI make sense to you and what shortcuts do I need to build into it for ease-of-use.

I am expecting that the early testers will uncover some bugs and will have some advice for ways to (re)organize the APIs, the simple GUI and/or the instructions for how to use the classes. I have had my head buried in this for so long, some things that seem obvious to me will probably be as clear as mud to you.

Once the code is ready for prime-time, users will be able to use it to write their own clients (AI or 3D graphics practice anyone?) and user interfaces, to develop simulations of sensors, to do some off-season experimenting with driver control hardware, etc. without having to implement the boring guts of client-server communication protocol.

If our physics were better and a library of accurate sensor simulations already existed, I would even suggest doing some autonomous control coding experiments, but I'm going to soft-sell that topic until the physics modeling gets a major refresh.

Blake

[Chris is me]

I'd be willing to help out.

[gblake]

Chris - Great - I'll take you up on it - If we are successful enough for you to be able to take a customized client to IRI that would be a nice achievement. PM me (or write to the address in my signature) to tell me how I can connect with you.

Brendan - Did I scare you off?

About IRI - I think that the IRI organizers are planning to repeat having 5th Gear there for the 3rd year in a row; but... Let's let them be the people who actually commit (or don't) - I think that they will be able to do it; but I'm not promising for them.

The other recent suggestions are all good (fix tunnler strafing, revive previous games, etc.) and are either problems we know we need to fix or improvements we hope to work in over the summer.

Cleaning up the existing code and so that we can efficiently expand the number people in our development team, and actually coordinating that expansion are competing for our time. The result will be good and will accelerate progress, but the start-up delay is frustrating for you and us - Thanks for sticking with us and for all the kind words about liking and using the simulator.

Let's get these client classes ready for distribution and then while you (our users) are having fun with them, I/we will move on to the next items in our fix/improvment lists.

Blake

[BrendanB]

Okay, I'm in! Sorry I've had a ton going on with getting ready for a vacation/roadtrip without robots involved !

[gblake]

Hey Folks - I just sent a (pre-Alpha level) chunk of code to my two Beta Testers - I/we will be in touch as the code, API specs, interface specs and any other documentation progress from today's levels to a ready-for-primetime release - Blake

[gblake]

Folks,

This is a new subtopic.

Imagine I am letting hordes of non-FIRST newbies use 5th Gear in an open-to-the-public exhibit. I am also letting some subset of the 5th Gear users drive a real FRC Breakaway bot after they use the 5th Gear simulator.

I want to use some non-arbitrary method for selecting which one of the six 5th Gear players gets to be the one person who earns some time at the sticks of the real robot.

I am thinking of inserting some sort of skill/performance metric tracker into the 5th Gear software. It would become sort of an automated scouting system that reaches a conclusion at the end of just one match and that relies on things that can be more or less easily tracked within the simulator.

Here is the result of my initial brainstorming (about 10 minutes of thought has gone into it so far - better ideas are welcome).

• Passing a ball from the defensive zone to the midfield = 1 scouting pt for the shooter
• Passing a ball from the defensive zone to the offensive zone = 1.5 scouting pt for the shooter
• Passing a ball from the midfield zone to the offensive zone = 1 scouting pt for the shooter
• A shot on goal = 1 scouting pt for the shooter (add 0.5 pt if the shot also qualifies as a pass from another zone into the offensive zone)
• A goal from anywhere = 2 scouting pt for the last friendly player that touched it (and maybe 1 pt for an ally that passed it to the shooter?)
• Blocking an opponent's shot on goal = 1 scouting pt

The player who accumulates the most scouting points in a match gets to drive the real robot for a couple of minutes.

Thoughts? Suggestions? Alternatives?

The results needs to reward good strategy and good execution from anywhere on the field, instead of only rewarding goals scored.

Blake

[JesseK]

As someone who's both driven the simulation and an FRC robot, I'd recommend constraining those who want to drive the robot to playing the simulation from the driver's station point of view only. That way the drivers who've never touched an FRC robot are less likely to ram it into something as soon as they touch the real robot's controls.

Does a midfield score = 2pts for person who shot it + 1pt for a "pass" from mid to offensive zone for a total of 3 points?

Does shooting a moving (kinetic) ball from defensive zone to midfield count as 2 points or 1? How should players realistically expect to get points for a 'blocked shot', given the extra computation needed to know if the ball was truly about to go in a goal?

Along the same lines, if someone miraculously scores a kinetic ball from the defensive zone, is that
1 + 1 + 1 + 2 = 4.5? (block + def->off + score)

[gblake]

Originally Posted by JesseK

As someone who's both driven the simulation and an FRC robot, ... less likely to ram it into something as soon as they touch the real robot's controls.

Good point (and we need to make the default sim's driver's station perspective more accurately represent standing in a real driver's station). The exhibit probably would not have a real drivers station, but it might... If the exhibit comes to pass, then for the duration of the exhibit we probably should have the player's viewpoint match the viewpoint they will have using a real bot (unless the real viewpoint is just too screwy for use in the simulated match).

Originally Posted by JesseK

Does a midfield score = 2pts for person who shot it + 1pt for a "pass" from mid to offensive zone for a total of 3 points?
....
Along the same lines, if someone miraculously scores a kinetic ball from the defensive zone, is that
1 + 1 + 1 + 2 = 4.5? (block + def->off + score)

Two answers:
1) All good questions to be answered before a system like this gets used.
2) I dunno! I literally spent no more than 10 minutes on the whole idea and most of those were spent writing and editing the post.

Assuming some method of picking one "winner" needs to be implemented can you think of any better method than a tuned set of heuristics that track moving balls toward and into goals?

Does anyone think any other criterion is better, or that should also be tracked?

Blake

[Chris is me]

I haven't done a lot of analysis of said criterion, but the real challenge is making sure that you're not biasing the results toward a particular field position, so that people fight over / for it.

[Jeff 801]

I am looking into using a Logitech dual action gamepad which is very similar to an Xbox controlers with 5th gear and I have found out that in most games that allow the use of xbox controllers there is a file called "xinput1_3.dll" and if you replace this file with one found on this link you can use just about any gamepad. I have looked in the application files directory and I cant seem to find it so wondering if I am not looking in the right place for it or does it even exist?

[gblake]

Originally Posted by Jeff 801

I am looking into using a Logitech dual action gamepad which is very similar to an Xbox controlers with 5th gear and I have found out that in most games that allow the use of xbox controllers there is a file called "xinput1_3.dll" and if you replace this file with one found on this link you can use just about any gamepad. I have looked in the application files directory and I cant seem to find it so wondering if I am not looking in the right place for it or does it even exist?

Jeff,

I passed your request to a colleague who should be better acquainted with the pertinent 5th Gear details.

I looked at my own 5th Gear installation directories and did not see any dlls. there can be more than one reason for this so that doesn't mean we/you are completely out of luck.

When we configure a Windows computer to run 5th Gear, we install a device driver that lets us talk to the handheld XBox controllers. Then, as far as I know we are done. So, unless the dll you listed will replace part of the device driver we install, I am a little pessimistic, but... Let's wait to hear from the other guy before throwing in the towel.

Blake

Let's let the other guy give us an answer before giving up.

[gblake]

FYI - Cross-posting a link to a related thread http://www.chiefdelphi.com/forums/sh...5&postcount=21