[robocup-msrs] Hello from the Microsoft Robotics Studio team

Wed May 16 13:57:14 EDT 2007

Hi Nikos,

1) Will we be given the interface definition for the new robudog drive, or we have to figure it out ourselves? The sample code currently implements a simulated differential drive which certainly not what we need.

Yes, you will be able to drive the robudog directly without having to use the simulated differential drive.  Robosoft is just finishing up a specialized soccer player service which allows more control over the movement sequences of the Robudog.  I will post another reply with a more specific schedule for when you can expect to see this sample service.  In the meantime, you can continue to use the Lego NXT soccer players to work on basic field movement and vision code.

2) Is it possible to get ground-truth information about the game state (robots/ball position) during the development phase? I noticed that there is a pose state variable for the player service which is however set to 0,0,0 by the referee. Can we possibly have a development-phase-referee assembly just for debugging?

The soccer referee currently doesn't allow this but we plan to release an update to the soccer package in about two weeks.  I will include this feature in the updated soccer referee service.

3) Is there any noise injected in motion/perception? If yes, shall we be given the characteristics of this noise?

We don't plan to inject noise in perception.  The only noise in motion comes from the Ageia physics engine and you will have to learn to compensate for that as you implement your soccer player services.

4) How can we reset robots/ball to specific positions during development?

The soccer referee doesn't currently provide an easy way to do this.  I will add this feature to the referee and it will be available in about two weeks.

5) Will there be changes to the code, or this is the version we are going to use in Atlanta? (clearly this is a very important issue given the little time until July).

I've already mentioned two planned changes:  Robosoft will provide a soccer player service that allows more direct control of the Robudog and Microsoft will ship an update to the soccer package in about two weeks.  There is also a possibility of additional changes if critical bugs or missing functionality is discovered.  Other than these changes, the software that you have is essentially the same as what will be running during the competition.

6) How can we guarantee that testing on a slow machine and running on a fast machine in Atlanta will not produce different results? What are the MSRS primitives that allow a universal (CPU speed independent) measurement of time?

This question deserves a much more in-depth discussion than what I can give right now.  All communication between the referee and players is asynchronous.  A slower referee machine means that the simulation will run at fewer frames per second and the physics engine will not be quite as accurate.  The player services will receive fewer updates and frames from the referee.  All data sent from the referee will be time-stamped according to the referee time.  I will follow up with additional information on this topic.

7) What are the similarities/differences of robudog with Aibo concerning kinematics?

I'm not familiar enough with the Aibo to be able to comment on this question.  I will encourage a Robosoft representative to post more information about the characteristics of the robudog.

-Kyle

________________________________
From: nikos.vlassis at gmail.com [mailto:nikos.vlassis at gmail.com] On Behalf Of Nikos Vlassis
Sent: Wednesday, May 16, 2007 1:22 AM
To: Kyle Johns
Cc: robocup-msrs at cc.gatech.edu
Subject: Re: [robocup-msrs] Hello from the Microsoft Robotics Studio team

Dear Kyle,

thanks for participating in the list. There are undoubtedly several issues that need clarifications. For instance: