Hi,
I just found out about Formula Pi a week ago and will probably give it a try to compete in the summer 2018 races.
To get a feeling for coding for Formula Pi, I have taken the code bases of 'monster', 'S2017-grand-pi-winner' and 'S2017-monster-winner'.
I've managed to get the python variants running, on my Windows 10 laptop, with the Simulator.
For the winning Monster code of last summer, which is written in java, I can't seem to find the 'SimulationFull.py' equivalent.
In the 'Formula.sh' file, the last line, which is commented out, looks like it is meant to be used to start in simulation mode.
When I try to start like that, the 'de.webtuples.formulapi.SimulationFull' can't be found.
However, I can't find that class or the java source in the code I have available.
What can I do to get this code started in simulation mode?
Kind regards,
Iwan
Useless comment
Hum, interresting, I think they didn't publish SimulationFull.java..... is it regular ?
The code on SourceForge is exactly as submitted for the final :)
I am guessing they stripped out anything unnecessary to keep things simple.
I have sent the lambda.p.racing team an email to see if they still have the file and are happy to share it.
//reboot while race ongoing?
File raceStartedFile = new File("/home/pi/formulapi/logs/raceStarted");
if (raceStartedFile.exists()){
logger.info("raceStarted-file detected");
Globals.setImageMode(ImageMode.FOLLOW_TRACK);
}
else {
Globals.setImageMode(WAIT_FOR_LIGHTS);
}
Thanks. :) Was necessary...
Thanks! I hope we don't need it again, but in case we are restarted again I'm adding it.
Cheers!
I haven’t provided the SimulationFull Class because I have stopped using it after Race 3 I think. But I’ve mailed the latest version to Arron today.
Summer 2017:
Race 1: Python (Java not ready)
Race 2, Race 3: Java, but massive problems (see videos). I haven’t seen these problems in the simulation
Rest of season: Java
Winter 2018:
Challenge 1, Race 1: Python
My test setup after Race 3 (Summer 17): small method based tests and the final test with cam, PicoBorg Reverse and monitor playing the image-sequence video.
Further problems with the simulation:
- no collisions possible
- track is too clean. My stuck detection detects false positives
- reimplementation of FLIPPED not testable in simulation
- problem with current track position. There is a big gap between the simulation values and the real track (till now I didn’t investigate on this). Plan is to replace it completely in Summer 18.
I will provide slides in about 2 weeks with more information.
Hi Thomas,
thank you very much for getting back so quickly with the SimulationFull class and the explanation of the status of the code.
Kind regards,
Iwan
I have checked the source files Thomas sent me into the SourceForge branch. If you checked out the code using git you should be able to grab the extra files using
git pull.I have had a brief look at the source and it seems easy enough too understand. Remember to correct the IP address on this line in
SiimulationFull.java:Hopefully it will work as-is, but we have been warned that it has not been tested with that version of the code and may need some small changes :)
When I have some time, I will see if I can get the simulation going.
Thank you Arron.
Add new comment