Francesco Ferro

Tracking objects in cluttered, contact-heavy environments is tough… especially when vision alone isn’t enough: V-HOP changes that by combining vision and touch for precise, real-time object tracking. Is a new visuo-haptic transformer that fuses visual and touch data, making object tracking more robust even when vision fails. Inspired by how humans use multiple senses, this method ensures stability and accuracy in real-world tasks. Why V-HOP stands out: ✅ Tracks objects even when vision is blocked or contact forces change ✅ Uses a unified haptic representation for better generalization across grippers and sensors ✅ Achieves smoother, more stable tracking by processing sequences instead of single frames ✅ Outperforms vision-only methods and enables precise robot manipulation in real-world tasks It proves that combining vision and touch is the key to reliable object tracking in robotics. Project: https://lnkd.in/dbgsrA7X Paper: https://lnkd.in/deyvz-z9 Thanks for bringing this to my attention, Hongyu Li 🫶

398

9 comentarios

Teaching by showing! 👨🏼‍🏫 Nordbo Robotics is making robotics accessible to all skill levels. Mimic is an intuitive tool allowing users to record and automate tasks without programming experience. By integrating Mimic with robotic systems, companies save time as operators can directly record movements, eliminating the need for extensive robot programming. They various industrial tasks like sanding, polishing, painting, and grinding, offering automation solutions for tasks traditionally requiring skilled human labor. 🧽 As you can see Universal Robots arm recreated the demonstrated path very accurately. Users can program Mimic-enabled robots with flexible options, including point-to-point movements, path generation, or demonstration. P.S. First deburring attempt by myself! ~~~ ♻️ Repost to help 1 robot find a new workplace.

1425

51 comentarios

Best generic robotics simulators? 🤖️ Simulation is one of the main ingredients in robotics development. But which simulator should you use for your robotics project? If you're still deciding, here are some of the top alternatives to generic simulators available in 2025: 🔹 O3DE – Originally a game engine, now, thanks to the people of Robotec.ai supports ROS 2 integration. - PROS: Realistic simulations are possible, thanks to the power of a game engine, including large multi-robot simulations - CONS: Still maturing, and a little bit complex to create ROS2 simulations. Installation is tricky. 🔹Webots – A beautiful and mature open source simulator with GPU support. - PROS: Very easy to use, many robots, worlds, and sensors available off the shelf. Very mature project - CONS: The integration with ROS 2 is a little bit tricky. It is a little bit slow if you don't use a GPU 🔹Gazebo Sim (Ignition) – this is the simulator that installs by default when you install ROS. The most widely used among the ROS community - PROS: Its smooth integration with ROS2. A good support for plugins that allow you to do virtually everything - CONS: Has many bugs and a slow development cycle. Still not possible to do proper grasps based on contact physics 🔹 Isaac Sim – Nvidia's robotics simulator brings photorealistic rendering and seamless integration with GPU-accelerated AI workflows - PROS: very powerful for reinforcement learning for robotics, its main application at present. The best simulator for perception simulations. - CONS: extremely complicated to install and use. It also requires large computer resources
 🔹 MuJoCo – Known for physics accuracy and performance, especially in manipulation and control tasks, widely used for reinforcement learning for robot control. - PROS: the best in physics simulation, like simulating contacts, complex structures interactions, and full robot body control - CONS: poor support of ROS 2 and almost non-existent support for sensors and full robotics environments And for the brave, try something different for fun! 🔹Minecraft with ROS2! Check out how you can connect Minecraft to your ROS 2 project with just two Docker commands: https://ow.ly/aj2G50WS2hV

 🔹 Roblox Studio: A free game engine where you can build interactive worlds — with aerodynamic forces supported to simulate drones: https://lnkd.in/dVSfiamA 💡Anyway, if you are starting in robotics, your best bet would be to go with Gazebo. Learn how to use it in our two special courses: 💻 How to use Gazebo Sim: https://ow.ly/ojX950WS2i4 💻 How to use Blender to create Gazebo Sim scenarios: https://ow.ly/mucM50WS2hW

226

21 comentarios

Real-time robotics can be hard — so here's another tutorial for doing real-time TCP/IP communication with ABB Robotics robots: https://lnkd.in/eisuBr-M Side note: Would a video tutorial be more helpful, or is the repo clear enough? Let me know in the comments...

63

3 comentarios

#ICSR + #ART 2026—Brilliant! It's fantastic to see more mainstream robotics conferences embracing art in such a major way. Well done to everyone involved, especially Heather Knight and Piotr Mirowski for this cool co-branding of #ICSR (https://icsr2026.uk/) with art! Art is increasingly being discussed within major robotics conferences as an important discipline rather than an afterthought. This is so reassuring for new researchers embracing this strongly interdisciplinary, yet precariously under-supported, emerging field. I hope the momentum continues and we build on a solid foundation, not a passing fad. This reminds me of the first fully integrated robotic art program at ICSR way back in 2014—more than a decade ago (thanks to Mary-Anne Williams!). The program included the workshop, "Misbehaving Machines," and the premiere of . Stelarc's magnificent 'SPINNING/SCREAMING: EVENT FOR THE AMPLIFIED HEAD.' It was a fabulous opportunity to see the master in his element. I still recall how he persisted in ensuring the projection was perfect, spending over a week at the Powerhouse Museum perfecting the shadow box, and then the entire day of the performance fine-tuning and rehearsing. Stelarc always speaks of the risk of trivialising art with technology. He warns of engineers creating "bad art." For those engineers wondering how sticking a banana on a gallery wall could or could not be art, reading Stelarc's body of work provides a glimpse into what makes art truly exceptional. Looking forward to this next instalment of ICSR + ART a decade on! https://lnkd.in/gD9AjiTD #RoboticArt #Stelarc #SocialRobotics #HumanRobotInteraction

58

2 comentarios

Now, anyone can build a bipedal robot with parts off Amazon. Open-source bipedal robot MEVITA, by Kento Kawaharazuka, is a platform with a minimal number of parts, all of which can be procured through common e-commerce sites. This approach, combined with the full release of its hardware, software, and learning environment as open source, empowers a global community of developers, students, and enthusiasts to not only build their own and contribute to its future. Link in to github in comments Follow Bots n' Beans for daily robotics, humanoids, and tech insights.

92

7 comentarios

🚀 Siemens Korea x Narnia Labs: Redefining Design with Generative AI We’re excited to unveil a groundbreaking collaboration between Siemens Korea and Narnia Labs — pushing the boundaries of design automation. Through Siemens NX and generative AI, engineers can input target performance parameters for a refrigerator duct and automatically generate optimized 3D models. These are seamlessly fitted into existing assemblies, performance-validated, and compared against legacy designs — all in one streamlined flow. This is what happens when advanced CAD software meets powerful AI agents. The design process becomes faster, smarter, and radically more efficient. 🔧 Goodbye manual iterations. 🤖 Hello AI-powered design. #SiemensKorea #NarniaLabs #GenerativeAI #NX #CAD

53

5 comentarios

Fun research project out of Shandong University From their abstract: "We designed a bipedal robot with quadrotor-assisted locomotion. The objective is to enhance the robot's motion performance through quadrotor assistance rather than achieving multimodal locomotion. In the prototype design, we improved the rigidity and compactness of the knee joint actuator by modifying the cantilever structure of the planetary carrier in the reducer to a bridge-like design. A simple motion control strategy was then developed to enable the robot to perform standing, walking, and jumping motions. Experimental results demonstrate that quadrotor assistance significantly improves both the stability and motion performance of the bipedal robot." Read the paper here: https://lnkd.in/e_EB7ERP

57

1 comentario