RoboComp is an open-source Robotics framework providing the tools to create and modify software components that communicate through public interfaces. Components may require, subscribe, implement or publish interfaces in a seamless way. Building new components is done using two domain specific languages, IDSL and CDSL. With IDSL you define an interface and with CDSL you specify how the component will communicate with the world. With this information, a code generator creates C++ and/or Python sources, based on CMake, that compile and execute flawlessly. When some of these features have to be changed, the component can be easily regenerated and all the user specific code is preserved thanks to a simple inheritance mechanism.

Cog and PyParsing

Let’s go to know how RoboCompDSL works.


One of the most important modules used in RoboCompDSL is pyparsing. It is a Python module whose latest version to date is 2.0.4, which lets you create grammars in a simple way. Thanks to this, we can create Parsers that are in charge of a grammatical analysis of files that follow a grammatical structure. This module is vital in generating the code as it is used to interpret the DSLs defined in RoboComp.

A simple example of it would be:

from pyparsing import Word, alphas
greet = Word( alphas ) + "," + Word( alphas ) + "!"
hello = "Hello, World!"
print (hello, "->", greet.parseString( hello ))


$ Hello, World! -> ['Hello', ',', 'World', '!']

This module is used in scripts and to extract the files CDSL and IDSL necessary information that will be used to generate code routine.


Cog is a code generator written in Python that allows you to inject Python code within a file so that, after using the generator, we get a result that depends on the injected code. It can be used with any type of file so it does not limit its use to code generation. The RoboCompDSL tool uses this code generator for creating and modifying components of a quick and easy basis of information obtained from CDSL and IDSL files. This tool is able to generate code routine that hindered the creation of complex systems components.

An example of use is shown in a file either:

Static part of our file
import cog
fnames = ['DoSomething', 'DoAnotherThing', 'DoLastThing']
for fn in fnames:
    cog.outl("void %s();" % fn)
Another fragment of our file

Running Cog:

static part of the file
void DoSomething();
void DoAnotherThing();
void DoLastThing();
Another fragment of our file

Update State Machine Code Generation in C++

Update State Machine Code Generation in C++

As “void compute;” no longer necessary, now when a component with state machine is generated the portion of “compute” no exist.

This method can be supress because Its funtion is realized by the state machine.

This portion disappears of files generated:

  • genericworker.h
  • genericworker.cpp
  • specificworker.h
  • specificworker.cpp

Progress with RoboCompDSL

Hi! I am back. It’s been weeks since my last post, because there was not much to say. In this post I will summarize the progress with RoboCompDSL. Now I finished the biggest work, we can communicate two Robocomp components using ROS middleware or ICE middleware and it is as simple as has been done so far. We just need write an IDSL(s) with some restrictions and a CDSL for our component.

What restrictions? Services has two parameters (in/out). Example: void service(int numRequest, out int numResponse); Topics has just one parameter. Example: void topic(int numToPublish); Allowed Types are structs, sequences(array) and primitives types as string, int, float…

Paying attention to this restrictions we can write an IDSL compatible with ROS and ICE middleware for our components. I am working on a guide to explain all the news and how to use them.

Finally I would like to highlight a further set of additions that I made:

Written a Syntax Highlighting File that allow in KDE applications results like:

Support QT5 completed and tested (an example of use in the image above).

Creating test for CNN object detection component in simulation enviornment (rcis)

Design Specifications:

The primary goal is to detect objects in simulation environment using CNN. This component gives a button based control panel to move omnirobot around. It also includes a button to grab image from RGB sensor and send to CNNobjectdetection module and get detections.

CSDL interface file for the component:

import "/robocomp/interfaces/IDSLs/RGBD.idsl";
import "/robocomp/interfaces/IDSLs/OmniRobot.idsl";
import "/robocomp/interfaces/IDSLs/objectDetectionCNN.idsl";

Component testODCNNsimulation
		requires RGBD;
		requires OmniRobot;
		requires objectDetectionCNN;
	language Cpp;
	gui Qt(QWidget);

Specific worker compute method:

  1. grab RGB image from simulator camera


  1. show camera output

    imshow(“Display window”,src);

Specific worker constructor:

callback declaration when button pressed.

connect(pushButtonLeft, SIGNAL(clicked()), this, SLOT(moveleft()));
  connect(pushButtonRight, SIGNAL(clicked()), this, SLOT(moveright()));
  connect(pushButtonFront, SIGNAL(clicked()), this, SLOT(movefront()));
  connect(pushButtonBack, SIGNAL(clicked()), this, SLOT(moveback()));
  connect(pushButtonClock, SIGNAL(clicked()), this, SLOT(rotateclock()));
  connect(pushButtonAnticlock, SIGNAL(clicked()), this, SLOT(rotateanticlock()));
  connect(pushButtondetect, SIGNAL(clicked()), this, SLOT(detectobject()));

Callback methods:

void SpecificWorker::moveleft()
	///move robot to left and stop		
cout<<"move left"<<endl;

catch(const Ice::Exception& ex)
	cout << "[" << PROGRAM_NAME << "]: Error in  proxy (base). Waiting" << endl;
	cout << "[" << PROGRAM_NAME << "]: Motivo: " << endl << ex << endl;
} } Similarly other callback methods are defined for other button clicked events. The detect object object button is similar to testobjectdetect for real images execpt the image is coming from rgbdproxy.



Update –Week 6

Recently I was coding some slot functions of what happens when we click on the actions like save,Open,exit.. The old code helped me lot in this case.. In between I tried to put a function to select what kind of context menu should be displayed on screen when we right click on QGraphicsScene..I mean the context menu which appears when we right click on node and its background should be different.I made a silly mistake during that and spend 2 hour going through documentations to understand my mistake :D :D..

Now I will be going into drawing the graphic Tree.Since the new nodes are not circular in shape,We can’t distribute the other connected nodes evenly around a node.So I will have to use a new algorithm for that..Basically I will have to use the same spring force system..But putting some limitations on the node connection directions..Once these drawing have finished,I have to enter the communication part..

I haven’t started coding any communication part yet..I had already learned some Basics about the Zeroc Ice for understanding the robocomp framework in the beginning.So it helps me to build the tool in complete independent modules for GUI,Communications and controlling components.

I have also decided to use subclasses of all three main classes in qt Graphics View now We are using the subclasses of QGraphicsView,QGraphicsScene and QGraphicsItem instead of using direct classes..This will help to add some extra functionalities to the tool in the future..

Basil M Varghese