Ilir Aliu's Linkedin Analytics

What if humanoid robots could moved with human-like balance, reach & flexibility? Tele-op using only VisionPro. ⬇️ Code & examples. AMO brings that dream closer to reality. (Adaptive Motion Optimization) helps humanoid robots control their entire body with precision, even during complex tasks like picking things off the floor. It mixes reinforcement learning and motion planning to deliver smooth, stable, and real-time control. Why AMO is a big step ✅ Combines sim-to-real learning with smart trajectory optimization ✅ Handles high DoF robots like the 29-joint Unitree G1 with ease ✅ Adapts in real time to new or unexpected commands ✅ Expands the robot’s reach and balance for real-world tasks It shows that whole-body control for humanoids isn’t just possible; it’s becoming practical. A solid step toward robots that truly move like us. Thank you, Xuxin Cheng for sharing! 🔗 Code & examples: https://lnkd.in/eyqb9Ppd 🌐 Project: https://lnkd.in/ea_gj88M

Meet SAND·E, a WALL·E-inspired robot that spots, picks, and bins trash on sandy terrain: What happens when you mix YOLO, Mistral, ACT, and Whisper… and send them to the beach? No fancy pipelines, just practical hacks and solid engineering. ✅ Splits screen into 4 zones to track sand and steer clear ✅ Uses Mistral 24B to spot debris when YOLO fails ✅ Voice-controlled with Whisper, ElevenLabs, and Mistral Large ✅ Picks up trash with ACT policy and a Pi0 backup Built with LeKiwi from @LeRobotHF, SAND·E adapts every time the environment shifts… one frame at a time. They won 1st place at the largest Robotics Hackathon in Europe earlier on April: Hugging Face x Mistral AI (36 hours)! Thanks for sharing, Alba María Téllez Fernández!

London, we’re coming! I’ll be at Humanoids Summit 2025, and here’s what I’m excited about: Chatting with Reachy 2 by Pollen Robotics (recently by Hugging Face) in multiple languages. Also looking forward to hearing from Remi Cadene! Can’t wait to see the mimic guys, Nicolas Heess, Director of Robotics at Google DeepMind and of course Pras Velagapudi, CTO of Agility Robotics. London, May 29-30, 2025! Full lineup and more info: https://lnkd.in/dUaqnVzi