In October, I attended the ROSCon UK 2026 conference in Edinburgh to give a talk about my rover. The talks and workshops were valuable, but the most useful parts were the conversations and community: from the author of the localisation package checking my rover configuration to finding people interested in a Cambridge robotics hackathon.
Prepare the Viam Rover for reliable autonomous navigation by installing its remaining software, calibrating the wheel encoders, and generating a motor lookup table. These steps give the control system more accurate movement data, ready for future mapping and navigation tasks.
This post is part of a series about the Autonomously Exploring Viam Rover (see my previous blog post). The rover is my attempt to build an inexpensive rover that is easy to assemble, so that anyone can replicate it. In this series, I'm showing how both the hardware and the software for the rover work.
In this post, I walk through how I built an autonomous indoor mapping robot using the Viam Rover as a base, swapping out the webcam for an RealSense Depth Module D421 and using a Radxa X4 as the central compute (instead of the already-supported Raspberry Pi or Jetson Orin). The robot runs a full ROS 2 stack, including Visual SLAM for map building and Nav2 for autonomous navigation.
FOSDEM is a conference which is free to attend. This year's conference (2026) was held in Brussels, Belgium over two days. The talks were separated into tracks based on subject, and one of the tracks was Robotics & Simulation. I managed to see almost all of the talks, but the room was constantly full with a long queue - clearly, robotics is popular!
This post is a summary of the talks with links and my own thoughts. If you'd rather go direct to the source, go to the FOSDEM 2026 website (specifically, the Robotics & Simulation track page).
Computer Vision (CV) is when a machine processes visual data to perceive its environment, just as humans can. Robots using CV can operate on single images or continuous video data, which allows them to recognise objects, detect obstacles, track movements over time, and navigate in complex space. For instance, autonomous vehicles can use camera data for safety systems or even fully autonomous driving.
This post walks through building a simple ROS 2 node that detects objects in real time using a robot's webcam and the Ultralytics YOLO26 model. The purpose is to show how simple it is to deploy Ultralytics models on a robot using the ultralytics Python package, as well as walking through a simple ROS 2 node for CV.
While looking for guest speakers for the Cloud Robotics Working Group, I came across roslibrust, which is a package for building nodes in Rust that can communicate with ROS. Given how different the approach is from ros2rust, I decided to make a few simple nodes in both libraries to see how they compare. This post compares the two libraries by writing the same application in each.
This post is also available in video form. If you'd prefer to watch, click the link below:
This post shows how to deploy an EC2 Spot Instance on AWS capable of running O3DE, Ignition Gazebo, and NVIDIA Isaac Sim. We’ll use AWS CDK, Amazon S3, and automated start/stop scripts to create an on-demand GPU simulation environment that preserves project data while keeping compute and storage costs under control.
Discover how large language models can translate natural-language requests into commands a robot can execute. Using Claude and a robot simulated in Gazebo, this post demonstrates flexible robot control before exploring how the LLM selects tools, breaks down tasks, and communicates with the robot.
This post takes a look at the PID Controller, which is widely used in robotics for providing more accurate motions of any moving parts on a robot. We will go through some basic control theory to give a better understanding of what a PID controller is, then look at each of the P, I, and D terms to develop a more intuitive understanding of the controller. From there, you should be able to use the controller in your own projects.