RoNeX plus Matlab

MATLAB now supports direct integration with ROS! Of course one of the first things we did was test this with RoNeX, and happily, it worked great. In this section, we detail how to use RoNeX with MATLAB (R2013a or R2013b).

If you don’t already have it, ROS support for MATLAB can be downloaded here:

http://www.mathworks.co.uk/hardware-support/robot-operating-system.html

Requirements

In this example, we are running Matlab and all ROS nodes (including RoNeX nodes) on a single machine. Since RoNeX is not supported on Windows, the example was executed on a 64-bit Linux (we used Ubuntu) machine. However, it is possible to run Matlab on a Windows machine, whereas RoNeX stays on a Linux machine, and let them communicate with each other over network.

Getting Started

Inside the package, there is one pdf file (ros-io-package-getting-started-guide.pdf). Simply follow the instructions in it to install the MATLAB ROS I/O package.

In the pdf file, there is one example that explains the basic instructions for creating publishers and subscribers inside MATLAB. Here, we explain how to publish data outside MATLAB and receive it inside MATLAB.

Assuming that you just want to run MATLAB and other ROS nodes on your local machine, set these environment variables (for example in the .bashrc file):

$ export ROS_HOSTNAME=localhost
$ export ROS_MASTER_URI=http://localhost:11311

ROS_HOSTNAME affects the network address of a node, and it is the key environment variable for this example. The ROS_MASTER_URI is an important environment variable that tells a node where the master is. However, great care should be taken when using localhost in ROS_MASTER_URI, as that can lead to unintended behaviors with remotely launched nodes.

Example Setup

Make sure that you have installed MATLAB, ROS Hydro, RoNeX software on the Linux computer, and the RoNeX hardware is connected to the computer.

In the following MATLAB code example, we create a ROS node inside MATLAB and then connect it to an existing ROS master (outside MATLAB). The ROS node inside MATLAB will act as the subscriber, whereas the RoNeX software will act as the publisher. Note that the topic name is my_data and the message type is std_msgs/Int16MultiArray in this example.

%% Create a Subscriber.
% Create a new node and connect it to the master.
% Type help rosmatlab.node in MATLAB to get more info about it.
node = rosmatlab.node('my_data_listener_node','http://localhost:11311');

% Add a subscriber to a topic named my_data to the node to receive message of type std_msgs/Int16MultiArray.
subscriber = rosmatlab.subscriber('/my_data','std_msgs/Int16MultiArray',100,node);

% Set my_function to execute when a valid message is published.
subscriber.setOnNewMessageListeners({@my_function});

%% Keep the node alive for a while.
for i = 1:10000
    pause(0.1)
end

%% Remove the subscriber from the node.
node.removeSubscriber(subscriber);

%% Clean up workspace.
clear;

Once an message is sent by the RoNeX software and then received by the ROS node inside MATLAB, we extract the integers stored inside the array data, and output it to the drive. Note that the layout info inside message is not processed here.

function my_function(message)
mydata = message.getData()
savefile = 'mydata.mat';
save(savefile, 'mydata');
end