Zumi-rpc

Remote procedure call server and client for remote control of Zumi robots.

• Python API mirroring the Zumi API from remote
• allows compute intensive robot algorithms
• uses the RPYC remote procedure call lib for efficient and secure remote calls

Setup

On the Zumi and the Client PC

Install rpyc:

pip install rpyc

On Zumi

use ssh or the Jupyter terminal to get on a console

Clone this Git repo:

git clone https://github.com/keuperj/Zumi-rpc.gi

Generate Certificates

in the Zumi-rpc folder, generate server credentials

openssl genrsa -out zumi_rpc.key 4096
openssl req -new -x509 -days 3650 -key zumi_rpc.key -out zumi_rpc.crt
chmod 600 zumi_rpc.key
chmod 600 zumi_rpc.crt

Then generate client credentials

openssl genrsa -out client.key 4096
openssl req -new -key client.key -out client.csr  -passin pass:""
openssl x509 -req -days 3652 -in client.csr -CA zumi_rpc.crt -CAkey zumi_rpc.key -set_serial 01 -out client.crt
rm client.csr

Now copy the client key and cert to your client machine and set proper rights:

chmod 700 client.*

Start Server

Use ssh or the Jupyter terminal to

python3 RPC-Server.py &

For an automatic start at boot time, you can add this command to /etc/rc.local

Working with Jupyter

The RPC Server does not interfer with the Jupyter frontend. You can run them both at the same time. Our typical workflow is to start the RPC-Server from a Jupyter terminal.

Remote access over the Internet

If you want to access the Zumi not only from it's own Wifi, but accross the internet, you need a relay server. Details are explained in our Zumi Repo - just replace the Jupyter Port 8888 with your RPC-Server port...

API

Zumi-RPC simply reproduces the original Zumi API, with some exceptions:

Camera

We bypass the Zumi camera methods and interface picamera directly, allowing us to get high resolution images:

get_picture(resolution=(1024,768))

returns a NumPy array of size (resolution_x, resolution_y,3) -> RGB image.

get return values

if you call a function which returns an object, you need to make a copy of this object. Otherwise, you will only have a local reference to this object and the data remains on the Zumi! So, for the cam example:

frame = np.array(get_picture())

will produce a local copy of the output.

Multtiple Connectons

The Zumi API and the RPC Server are non-blocking. Hence, you can have multiple RPC-clients connected to one server as well as use RPC and Jupyter simultaneously.

More Zumi Libs and Hacks

see our Zumi Repo