Kshitij Tiwari Ph.D.

10 Alternatives to ROS for robots

7 min read

Robot Operating System (ROS) is the most commonly used platform for robot development. It is an open-source robotics framework that provides all the necessary tools and features for development, including libraries, drivers, visualizers, building tools and more. The open-source nature allows for reuse of some of the software architectures published by peers in the community to avoid having to reinvent the wheel over and over.

However, it is not the only option for robotics development and a number of alternative frameworks exist which could be better suited to different types of projects and applications.

In fact, there are some robotics companies such as Pulu Robotics, that do not use ROS since the inception and for various reasons. Now, without getting in the analysis of their reasoning for avoiding ROS, let us look at some of the most popular alternatives to ROS and discuss their pros and cons to help you decide for yourself whether you need to integrate ROS in your workflow or avoid it.

When to consider using ROS?

ROS is a powerful platform for robotics development and there are a number of situations in which it makes sense to use it. It is a good option for developers who are familiar with robotics and have complicated projects with real-time demands.

ROS has excellent support for multiple sensors, cameras and other peripherals, and is particularly useful for projects which require any kind of computer vision. ROS also performs well with SLAM (simultaneous localization and mapping) and can be used for autonomous vehicles and robots with precise navigation.

When to consider alternatives to ROS?

ROS is usually perceived as somewhat mandatory for robotic development though it may not always the best option for robot development. It can be complex to learn and implement, and is not suitable for simpler projects.

Here are some of the scenarios when considering alternatives to ROS would serve the needs better:

  • Time crunch and steep learning curve: If time is of the essence, alternative development frameworks might be better suited, as ROS has a slightly longer development cycle. There are also some projects which do not require the level of complexity that ROS offers and sometimes, for business needs, ROS tends to overcomplicate things.

  • Importing custom robot models in ROS: If you are using a custom robot, depending on the complexity of the design, it might become a hassle to import the robot model inside ROS. While the CAD tools like SolidWorks allow exporting 3D models to URDF, other CAD tools might not have this option.

    If you didn’t know this already, URDF stands for Unified Robotics Description Format, is an XML specification which is required for a 3D model to be integrated into the ROS framework.

  • Performance and functional limitations: For commercial applications, software development with ROS framework may not meet quality control standards or might require additional work to get there. There might also be some application-specific feature which might be desired but might not be natively available within ROS.

  • Platform lock-in: Developing the robot software stock within ROS basically locks you in. In the future, if you wish to move away from ROS or use an alternative framework, that might result in significant overheads.

  • Latency in message passing and decision making: ROS is a distributed control architecture which has many pros such as distributing the computational overheads but it also can become a bottleneck.

    Message passing may be slowed down inducing latency in some cases and needs to be handled by modifying robot behaviors such as moving slower than intended to allow time for messages to arrive.

Alternatives to ROS for robots

There are a number of alternatives to ROS depending on the specific needs of the project. Some of the most popular include:

  • PyRobot – this is an open-source robotics framework which is particularly well-suited for academic and research projects. PyRobot provides a unified API to control robots from different vendors and allows for easy development and deployment of experiments. Offers functionality for implementing manipulation, navigation and pick-and-place demonstrations.

  • Orca – this is a C++ framework for robotics development which is designed for embedded systems and low-power applications. It provides a software stack for controlling low-level peripherals, as well as support for higher-level control algorithms such as motion planning and machine learning.

  • Yet Another Robot Platform (YARP) – this is an open-source robot operating system (ROS), written in C++ and capable of running on Linux and Windows. It is flexible and extensible and provides a range of tools for automation, robotics, and machine vision.

  • Open Real-time Control System (OROCOS) – this is an open-source real-time control system (RTS) designed for robotics applications. It is written in C++ and targets the Linux platform and is capable of distributed execution with support for real-time scheduling.

  • genom3 – this is a distributed robotics control system which is designed for real-time, safety-critical robotics applications. It is written in C++ and provides a range of features for real-time scheduling, message passing, and distributed control.

  • Mobile Robot Programming Toolkit (MRPT) – this is an open-source C++ library for robotics applications. It provides features for mapping, localization, path planning, navigation, and computer vision and is easily extensible and scalable.

  • Integrated Development Environments (IDEs)– IDEs like NAOqi are suited for Aldebaran robots such as Pepper and NAO, while V-REP is suitable for a wide range of robots, including those running on ROS.


    Similarly, Webots is a powerful and popular robotics development environment designed to facilitate the development of robotics applications. It provides an integrated development environment (IDE) for writing and debugging code, a physics engine for simulating dynamic systems, and a visual programming language for creating control logic. Webots also includes libraries for a wide range of sensors, actuators and controllers, as well as a number of tools for 3D modeling.

  • Player Project – this is an open-source frameworks for mobile robotics research and development which is capable of controlling a range of robots. It provides easy-to-use APIs for controlling a wide range of robots, as well as a number of other features such as obstacle avoidance and path planning.

     

While ROS has improved a lot with time and the open source nature with an active community allows for reusing some of the solutions already developed by peers, using an alternative platform to ROS can often result in a faster development cycle. Additional, it can offer simpler integration, and more customization options for complicated projects. It is worth considering non-ROS options for robotics development where appropriate.

Key takeaways

Alternative frameworks exist for robotics development aside from Robot Operating System (ROS). When selecting which to employ, consider the features and capabilities that you need for your application and the associated development cycle. Research thoroughly to decide which is optimal for your use and understand the pros and cons of each approach before settling on ROS or alternatives to ROS for your end use.

Related resources