Using Robo HAT MM1 with Donkey Car

I am using a Robo HAT MM1 with my Rpi 3B+ on my Donkey Car.

I am following these instructions https://www.hackster.io/wallarug/autonomous-cars-with-robo-hat-mm1-8d0e65 to install the Robo HAT MM1 on my Rpi 3B+ and then program both the Robo HAT MM1 and the Rpi 3B+ with the appropriate development software.

I have installed the “rc-cpy-final.py” program on the Robo HAT MM1 and have changed it to “code.py” as per the hackster.io instructions.

The last thing I need to do is install the Donkey Car software on the Rpi 3B+. To do so at Step 11 of the Donkey Car Rpi software installation I believe I should change “git checkout master” to “git checkout dev” to install the Donkey Car software that will work with the Robo HAT MM1? Is this correct?

Regards,
TCIII

Hi, yep that‘s right. Saw the robohat.py part only on the dev branch which is default now. robohat works perfectly with the jetson nano and I even got the imu recordings just with doing the setup in the myconfig.py.

@Rainer,

Thanks for the info, much appreciated.

If I go to github.com/autorope/donkeycar I get the dev branch, yet in the Step 11 Donkey Car software installation, it still says “git checkout master”.

Are you saying that the “dev branch is the default now” as there is still a “master” branch in the drop down.

I am presently running a Jetson Nano on my Donkey Car chassis, but I plan to switch to the Rpi 3B+ because I don’t like hanging the Robo HAT off the end of the Nano GPIO bus.

How did you support the Robo HAT on the Nano?

Regards,
TCIII

I just tested the recording with the software. You could just use a cable.

As shown, the dev branch of donkey’s github repo is the default. The documentation is for the stable branch which is master. So if someone tells you that the software part you need is in the dev branch, then you don’t have to follow the step in the documentation which has you getting the master branch…

Be aware though, it is a branch under development( hence the name ) and something might break so you will want to keep one sleepy eye on the github repo to see when updates to the repository happen so you can put 1+1 together if your software all of a sudden stops working.

Also, one of the things I’ve done in the past is to have customized parts of the master repo(locally of course) to use things which were in dev but with the stability of the master branch. Some donkey parts are pretty much self contained so the change to master is only within the one parts file and then your own manage.py file.

But this does take a basic understanding of python and coding to implement and there are no “instructions” on what or how it’s done.

Hi All,

After installing Buster lite instead of Stretch lite I also installed pip3 and both tensorflow and OpenCV installations and importing worked fine.

The “donkey creatcar --path ~/mycar” did create a “mycar” in the root directory finally unlike when using Stretch lite.

However the config.py and myconfig.py do not contain “HAVE_ROBOHAT” unfortunately.

I tried using both “git checkout master” and “git checkout dev”. When I used “git checkout dev” I received a message that dev and master were merged, but obviously not!

Can I remove the “donkeycar” and “mycar” directories and start over with the correct dev software or is there a way to scrub both that does not require removal of the directories?

Regards,
TCIII

after you update or switch your branch you just update donkey( pip install -e .[rpi] ) and you must recreate your base mycar directory.

this should be stated in the docs. is it not?

Hi DougL,

Keeping things up to date.

I assume that you would change “git pull” to “git checkout dev” which I have already tried?

“cd projects/donkeycar
git pull
donkey createcar --path ~/mycar --overwrite”

Running:

“cd projects/donkeycar
git pull
donkey createcar --path ~/mycar --overwrite”

Comes back with “everything is up to date”

Regards,
TCIII

first you do what you want to the repository( switch branch, update, etc )
secondly you update your local python environment with the updated software( pip install -e . )
last you recreate your donkeycar runtime with donkey createcar …

If you never update your python environment then there will never be a change to your runtime. The only exception is when you personally edit one of the python files in your donkey repository directory.