Beginner’s Guide to ROS — Part 4 “Services In-Depth” Tutorial

Krishna Shah
6 min readApr 1, 2022

Who are we serving again?

Recap: What is a Service

Remember from the first part of this series, services are a node that allows you to give and input to get the desired output. If we want to add values, we can make a service that takes two numbers, sums them, and responds with the sum. Remember, they don’t publish continuously. They only respond when the service is called.

All good? Time to serve some nodes!

Format of this and future articles:

It is important to understand how this article will go. We will go through an example together. I believe that knowledge can only be attained through practical examples.

Task 1

Take in 2 numbers and respond with the product of those numbers.


A node will programmatically call the service and we must do the following:

  • Get inputs
  • Compute product
  • Respond via topic the product


Making a ROS Package

First, we have to make a ROS package which will be our node. That node will contain the code to make a topic and publish the message. There is a built-in command to do this using catkin.

First, though, navigate to the workspace:

cd catkin_ws/src

NOTE: You MUST be in src.

Then run this:

catkin_create_pkg basic_ros_service std_msgs rospy roscpp --pkg_version 0.0.1 --description "Basic ROS Service" --license MIT --author "Krishna" --maintainer "Krishna" --rosdistro "noetic"

Yikes! That is a lot so let me break it down.

  • First, we have catkin_create_pkg which is the base command.
  • Next comes the name of the package, in this case basic_ros_service . This can be anything.
  • Then, we add all the dependencies of the package. Here, we have 3: std_msgs rospy roscpp
  • stg_msgs is a ROS package that contains a bunch of data types like Float64, String, Int64, and more!
  • rospy is a library that contains methods for using ROS with Python. This is what we will use today.
  • roscpp is a library that contains methods for using ROS with C++. We will NOT use this today.
  • rosdistro states the distribution of ROS we are using. In this case, it is ROS “noetic”. (This is optional as long as you have sourced your ROS environment)
  • — Note: Services have additional requirements that we will manually add later!

Everything that is mandatory to install has been done. Now on to the optional stuff.

  • These parts are really self-explanatory. It includes the license, description, author, maintainer, and version.
  • You don’t need this but it is a good practice to keep.

Now, list the files in the directory. You will see this:

basic_ros_service  CMakeLists.txt

Go to basic_ros_service and list again.

CMakeLists.txt  include  package.xml  src
  • CMakeLists.txt — contains all the info catkin needs to compile the project. It contains a list of the libraries, files, packages, and other stuff it has to compile. We will come back to it later.
  • package.xml — contains info like package owner, description, version, maintainer, license, etc. We filled some of it earlier. The only important thing in it is the dependencies definitions. Our command already did this previously.

It looks like this:

***A BUNCH OF OTHER STUFF***<!-- The *depend tags are used to specify dependencies -->
<!-- Dependencies can be catkin packages or system dependencies -->
<!-- Examples: -->
<!-- Use depend as a shortcut for packages that are both build and exec dependencies -->
<!-- <depend>roscpp</depend> -->
<!-- Note that this is equivalent to the following: -->
<!-- <build_depend>roscpp</build_depend> -->
<!-- <exec_depend>roscpp</exec_depend> -->
<!-- Use build_depend for packages you need at compile time: -->
<!-- <build_depend>message_generation</build_depend> -->
<!-- Use build_export_depend for packages you need in order to build against this package: -->
<!-- <build_export_depend>message_generation</build_export_depend> -->
<!-- Use buildtool_depend for build tool packages: -->
<!-- <buildtool_depend>catkin</buildtool_depend> -->
<!-- Use exec_depend for packages you need at runtime: -->
<!-- <exec_depend>message_runtime</exec_depend> -->
<!-- Use test_depend for packages you need only for testing: -->
<!-- <test_depend>gtest</test_depend> -->
<!-- Use doc_depend for packages you need only for building documentation: -->
<!-- <doc_depend>doxygen</doc_depend> -->
<!-- The export tag contains other, unspecified, tags -->
<!-- Other tools can request additional information be placed here -->

That bottom part with the buildtool_depend and build_depend statements define the important libraries.

First, we must make a folder in our package called srv/ . This will hold our custom service. In it make a file called Product.srv . Put this in it:

float64 a
float64 b
float64 product

The first 2 float statements are the input. Then we put --- to show ROS that we are now defining the output. Lastly, we have the output. This allows us to define custom services with custom inputs and outputs.

Navigate to the src within that. Now you should be in USERNAME/catkin_ws/src/PACKAGE_NAME/src .

Here is where we put all of our code! Make a file called . Again, names aren’t super important.

Here is your file structure right now:

- build
- devel
- src
- basic_ros_service
- CMakeLists.txt
- package.xml
- include {This directory contains nothing}
- CMakeLists.txt
- src
- srv
- Product.srv

Now let’s start on the code itself. First, you have to import the rospy and std_msgs .

#!/usr/bin/env python3from __future__ import print_functionfrom basic_ros_service.srv import Product, ProductResponse # Import the compiled version of our service. There aren't files called ProductResponse but ROS will understand that. 
import rospy
import math

Now we make the function that serves the clients.

def product_func(a, b): # Actual multiplication function
return a*b
def compute(req): # Find the product from the inputs req.a and req.b and return the product.
product = product_func(req.a, req.b)
print("Returning [%s]" % (product))
return ProductResponse(product)
def product_server(): # Start the node and server.
s = rospy.Service('basic_ros_service', Product, compute)
print("Ready to compute.")

And then we can run it with this:

if __name__ == "__main__":

IMPORTANT: You much make sure that your Python file is executable. sudo chmod +x

Go the CMakeLists.txt and add our python script to the file. This will make sure that catkin will compile it.

# include

Then, we need to make our SRV file that also needs to be compiled. Add the following lines to your CMAKELists.txt:

## Generate services in the 'srv' folder
## Generate added messages and services with any dependencies listed here

Also, add a dependency to the find_package section:

find_package(catkin REQUIRED COMPONENTS

We also need to add this dependency to our package.xml. Add the following lines next to all your other depend statements:


Now go back to catkin_ws . Run catkin_make . Source the directory source devel/setup.bash .

If it compiles, the hardest part is over!

Running it

Start roscore : roscore . Now, start serving: rosrun basic_ros_service

Check the node using rosnode list . We will see our basic_ros_service service.

Check the service using rosservice list . We will see our basic_ros_service service and some logger services. Let us try calling our service using CLI.

rosservice call /basic_ros_service "a: 2.0
b: 2.0"

Hint: If you forget the format of the command hit TAB twice and it will auto-populate the command.

We will see our result of 4.0 .

Receive Data Programmatically

Import libraries:

#!/usr/bin/env python3from __future__ import print_functionimport sys
import rospy
from basic_ros_service.srv import *

Define client function and make the appropriate call:

def product_client(a, b):
rospy.wait_for_service('basic_ros_service') # Find our service
solver = rospy.ServiceProxy('basic_ros_service', Product) # Connect to the service
resp1 = solver(a, b) # Make a request
return resp1.product # Return the result
except rospy.ServiceException as e:
print("Service call failed: %s"%e)

Make the usage helper in case we forget how to use it:

def usage():
return "%s [a b]"%sys.argv[0]

Run the functions from the CLI arguments:

if __name__ == "__main__":
if len(sys.argv) == 3:
a = float(sys.argv[1])
b = float(sys.argv[2])
print("Requesting %s*%s"%(a, b))
print("%s * %s = %s"%(a, b, product_client(a, b)))

Again, remember to change the permissions of the file. And run it the same way as the publisher.

To run it, do rosrun basic_ros_service 2 2


In this article, you learned how to make a Service Server and Client. You also learned how to make custom messages in the form of service. Lastly, you dived a little deeper into the CMakeLists and the message_generation library. If you encounter any errors or have any questions, leave a comment. I will try my best to help you. I hope you enjoyed it. If there is a particular topic in ROS that you want me to cover, let me know in the comments as well!

DeepShift Labs

At DeepShift Labs, we aim to make General Intelligence a reality. We have made significant progress towards this goal. If you are interested in what we do, please check out our Github and YouTube links! If you support General Intelligence and our project, please consider starring, forking, or contributing to our repository.

Thank you for your support!



Krishna Shah

I am a robotics engineer and AI enthusiast. I have made many cool robots and revolutionary AI algorithms.