Can I run MATLAB Code on a Raspberry Pi?

I'm planning to use a Raspberry Pi to interface with a sensor, and I’d like to control and monitor it using MATLAB. My goal is to perform calculations and visualizations based on the sensor data. However, I’m unsure whether these computations and visualizations can be handled directly on the Raspberry Pi, or if they need to be processed on my desktop, which is connected to the Pi. I have a few questions regarding this setup:

1. Can you install and run MATLAB on a Raspberry Pi?
2. If you cannot, how can I control the Sensor using the Raspberry Pi with MATLAB commands?

Answer 1

[MathWorks Support Team]

How to run MATLAB code on a Raspberry Pi:

While you cannot install the full MATLAB desktop environment directly on a Raspberry Pi, you can still run MATLAB code on it using one of the following two supported workflows:

1. Interactive Communication: In this workflow, MATLAB runs on your desktop or laptop, and communicates with the Raspberry Pi over a network connection. You can acquire data from sensors via GPIO, I2C, SPI, and Serial, camera modules, USB webcams, and more. This method is simpler to set up and ideal for prototyping or data acquisition tasks.
2. Standalone Execution: This workflow allows you to deploy MATLAB code as a standalone application that runs directly on the Raspberry Pi. However, it requires MATLAB Coder to be installed on your development machine. For more information and examples, see the following documentation page: https://www.mathworks.com/help/matlab/supportpkg/deploy-matlab-function-on-raspberry-pi-hardware.html 

Required Software

Both workflows require the MATLAB Support Package for Raspberry Pi Hardware to be installed on the development machine (not the Pi). This package enables communication between MATLAB and the Pi and provides functions for hardware interaction . You can install the package via the Add-On Explorer in MATLAB.

You can also use Raspberry Pi with MATLAB Online: Using a Raspberry Pi with MATLAB Online

[Walter Roberson]

Within possibility (1) there are two major distances in workflows.

In one workflow, requests are made to functions at the MATLAB level, and the MATLAB level communicates the requests to the Ardunio which branches to the appropriate section of code to take action and opaguely returns results back up to the MATLAB level.

In the other workflow, C or C++ code is deployed to the Arduino, with the Arduino potentially looping on its own emitting data to the serial port that is (eventually) read by reading the serial port at the MATLAB end. Possibly it is set up so that the MATLAB ends sends a request via the serial port for the arduino to take action on and return serial port data; possibly it is set up so the Arduino is free-running occasionally emitting something to the serial port.

For example the first workflow might involve calling arduino(), then device() to set up I2C, then read() from the I2C. The arduino monitor has to be ready for all kinds of commands, so the read() gets translated into a command sequence sent to the ardunio that gets decoded and analyzed and then branched to the appropriate section of code for action.

For example the second workflow might involve calling serialport(), then readline() -- with the code previously deployed to the arduino having been hard-wired to loop reading from the particular I2C port and println() the reading to the serial port.