Yasin Yousif — Mon, 09 Feb 2026 23:43:50 GMT

Robot Learning is the field utilizing Machine Learning and Deep Learning methods for training AI models to perform new physical skills on a robotic platform. The same term is also used for a new type of job titles, namely robot learning engineer, which is commonly focused on Imitation and Interactive learning of robotic skills from recorded datasets. This last formulation also features key words such as diffusion policy, or vision-action-language models. However, many other requirements also exist, such as robotic theory methods for navigation or arms control (like SLAM Filtering methods or Forward/Inverse Kinematic) as well as the usual software development skills like DevOps, Gitlab flow and Cloud computing. In this post we try to draw a full picture of these requirements, why and when they are needed. Additionally, suggested resources to study and projects to build for learning them are also mentioned along with an estimated time for doing that.

Read more… (14 min remaining to read)

Yasin Yousif — Mon, 01 Dec 2025 03:41:14 GMT

For any physically-intelligent mobile robot that operates autonomously in the real world, successful navigation is a critical capability. This requirement of the topic is also highly relevant across various sectors, including logistics, autonomous driving, and search-and-rescue applications. Furthermore, advanced robotic tasks, such as object manipulation, often depend on the foundational stability provided by accurate navigation. While the specific solution approach must adapt to environmental parameters (such as indoor/outdoor or known/unknown maps), all effective systems fundamentally rely on robust path planning. Based upon the foundational concepts of robotics, such as planning methods (Bug1 and Bug2), this guide provides a comprehensive, yet accessible, depiction of mobile robotics with a demonstrated 2D example. Using the popular ROS2 framework and C++, we will explore the core ideas needed to manage basic wheeled robot navigation in a simulated 2D environment to reach a goal while avoiding obstacles, explaining the theory as well as the code.

Robot Learning by Example (Posts about ROS2)