Here is your #ROS and open source robotics news for the week of June 8th, 2026: 🚀ROS Meetup @AutomateShow 🚀ROSCon Global 2026 talk schedule released 🚀New ROSCon Italy this November in Bologna 🚀@opencvlive 5 drops: better, faster, more AI 🚀@NASA Space Robotics Challenge ⬇️⬇️

Here is your #ROS and open source robotics news for the week of June 8th, 2026: 🚀ROS Meetup @AutomateShow 🚀ROSCon Global 2026 talk schedule released 🚀New ROSCon Italy this November in Bologna 🚀@opencvlive 5 drops: better, faster, more AI 🚀@NASA Space Robotics Challenge ⬇️⬇️

Another post @kamathsblog! In this blog, he published a detailed walkthrough on adding autonomous navigation to the @huggingface LeKiwi robot using Nav2 and ROS 2, visualized entirely through Foxglove. LeKiwi is an open-source holonomic platform, and this post covers the full stack from sensor prerequisites to sending navigation goals on a live map. The implementation includes LiDAR angular filtering with laser_filters to remove a pan-tilt camera obstruction, SLAM-based mapping with slam_toolbox in async mode, and a choice between two localization methods: slam_toolbox's built-in pose graph localization or AMCL with a particle filter. For navigation, Aditya went with an MPPI controller configured for an omnidirectional base and a SMAC 2D planner. He also wrote a small twist_switch_node to toggle cleanly between teleop and Nav2 velocity commands via a controller button — a neat practical touch. It's a good reference if you're setting up Nav2 on a non-differential drive platform. Blog post: hubs.li/Q04lcrkk0

#ROSConJP2026 のスポンサー様一覧を更新しました。 たくさんのサポートをありがとうございます！ roscon.ros.org/jp/2026/#spon… #rosjp #rosconjp #roscon #ros

LED Plugin has been merged upstream in @GazeboSim!🎉 This plugin makes it possible to simulate configurable LED systems in Gazebo. Do give it a try.

／ 📢 #ROSConJP2026 プログラム公開！ ＼ 2件の基調講演・20件の講演・1件のパネルトークからなる、#ROSConJP2026 のプログラムを公開しました。 ROS三昧の2日間になりそうです！ roscon.ros.org/jp/2026/#prog… #rosjp #rosconjp #roscon #ros

Well, would you look at that! fanucamerica.com/products/so…

Just shipped @rosorg based devcontainer support for the Artemis Mission Simulator 🌕! Check it out: Merged PR: github.com/jasmeet0915/artem… Repo: github.com/jasmeet0915/artem… The project now supports 2 development workflows: 1️⃣ One-click setup with VSCode's devcontainers. 2️⃣ Manually build and run the docker dev container using the helper scripts.

A big thank you to @TechCrunch for featuring Hello Robot and our mission to bring practical robotics into real homes. We're proud to be building technology that helps people maintain independence, dignity, and agency in their daily lives. 🔗techcrunch.com/2026/06/04/is…

Compact open source 3D printable mobile robot github.com/Obijuan/Miniskybo…

OpenCV 5 open-source computer vision library released cnx-software.com/2026/06/10/… The main change is the implementation of a new DNN engine with better ONNX coverage to enable a wider range of AI models for detection and segmentation, as well as vision language (VLM) and generative text (LLM) models. It runs on the CPU only for now, but GPU and non-CPU HAL support are being worked on. Other changes include Core changes for new data types and performance optimization, improved 3D vision support, and revamped documentation.

We at b»robotized just love Gaussian Splats! That is why we were super thrilled when we read that someone made a plugin for displaying splats in Rviz. So we had to try it right away! Video: Our #AID4SME project in Isaac Sim & Rviz, both showing the GS of the lab

学生時代、というか学部3年の頃から標準SLAMとCartographerを比較してみようって先輩に言われてたのに、のらりくらりしてしまいもう卒業してしまった今になって気になってしまいSLAM ToolboxとCartographerを比較してみました✋ 1つ目の動画がSLAM Toolboxで 2つ目の動画がCartographer #ROS2 #SLAM

ROSCon Globalに2件採択されました🙌🙌（採択率24%！） - Dynamic Window Pure Pursuit in Nav2 Regulated Pure Pursuit DWPP含むNav2 RPP controllerの各手法の特徴を説明します - ros2log new CLI and rcl_logging_implementation for Lyrical Luth ROS Japan dev会議で藤田さんと取り組んだ内容です

🤖 ✨ Following up on my earlier post on my 'Home Robot' with reactive obstacle avoidance navigation nearly a month ago, I am super excited to share my next technical milestone: proactive voice vision navigation assistant, integrated and all running on-device. Goal: Autonomous home assistant 🤖 (say Alexa on wheels 😉) Capabilities: - Completely vision based (the only sensor is an RGB-D camera). Maps the whole area and stores landmarks used for navigation. - Hands-free conversation: wake-word activation, then continuous follow-up dialogue with short-term memory; auto-sleep on silence. Voice-commanded autonomous navigation to landmarks. - Multimodal scene understanding: "what do you see?" answered by an on-device VLM from the camera feed - Real-time object detection via YOLOv8n Tech Stack / Tools (100% on-device): Master controller: @NVIDIAAI Jetson Orin Nano Super Chassis: Waveshare wave rover 4WD Audio: ReSpeaker XVF-3000 mic array (AEC/beamforming), Piper TTS Speech: openWakeWord whisper.cpp (CUDA / GPU accelerated) Language/Vision: @GoogleAI Gemma 4 E2B multimodal via llama.cpp (GPU accelerated) Perception: YOLOv8n on TensorRT (GPU accelerated) Autonomy: @OpenRoboticsOrg ROS2 Humble, Nav2 (Planner: Smac Hybrid, Controller: RPP), RTAB-MAP VSLAM (EKF fused RTAB-MAP's visual odometry with dead-reckoning odometry) Peak memory: ~6.7GB / 8GB with the full stack live Key engineering takeaways: - NVIDIA developer forums are great, they helped me solve a lot of issues really fast. - It took me a while to figure out the right VLSAM approach for my vision based sensor. The learnings here have been immense RTAB-Map for RGB-D camera, SLAM Toolbox for LiDAR based nav, CuVSLAM for Stereo camera - On constrained edge hardware, the system-level bottleneck is memory and GPU contention, not model quality. Right-sizing the models (2B over 4B, small over medium STT) was what made concurrent operation stable. - A tiered runtime: a lightweight always-on tier an on-demand navigation tier, was essential to verify everything fit in 8GB. - The bulk of the engineering effort was robustness: surviving audio glitches, boot-time race conditions, and resource pressure, not adding capabilities. Here's 🥂 to my technical PoC: a robot that sees, hears, reasons, and moves, with nothing leaving the device.