Edge Impulse’s Post

2024 is the year of Humanoid Robots. NVIDIA Joins the Race with Project GR00T! NVIDIA just unveiled their groundbreaking initiative to solve embodied AGI in the physical world which they have called "Project GR00T". At the heart of this project is GR00T (General Robotics 003), a foundation model for humanoid robot learning that understands multimodal instructions and performs various tasks. Jenson Huang has announced Project GR00T to be a cornerstone for the "Foundation Agent" roadmap of the newly founded GEAR Lab, where generally capable agents are being built to act skillfully in virtual and real worlds. With NVIDIA joining the race, 2024 might just be the year humanoid robots take center stage! #NVIDIA #ProjectGR00T #EmbodiedAGI #HumanoidRobots2024

Time to migrate your robot simulation! With Gazebo classic's EOL arriving in a few months' time, it's time to explore the various ROS2-compatible simulation possibilities for your projects. There are several alternatives: •  Gazebo Sim: A safe choice for teams used to Gazebo Classic and who want a similar experience (Migration Guide available). •  O3DE: An open-source 3D engine multipurpose that propose a ROS2 integration and is still growing in popularity. •  Isaac Sim: A robotic simulator developed by Nvidia, featuring photo realism and supplied with numerous Nvidia composites. There is also Webots and Mujoco but I haven't tested them yet, so I'd be interested in your feedback. Which simulator will you use for your ROS2 projects? If you would like to share your experiences with these simulators, or with others I haven't mentioned, let a comment 💡 Follow for more ROS2 and robotics content! Let’s build a stronger ROS2 community together 🚀 #ROS2 #Robotics #Community #Simulation

It is always fun to read comments or posts about Robotics simulator comparisons. I have been using many simulators over the last 8 years, and one thing has not changed. A general sense of confusion in the industry/academia when choosing a simulator for a Robotics application. The most obvious takeaways are: 1. There is no single best solution to fit all problems. Each domain has its simulator. 2. The current technological gap for decent simulators is a legitimate business case for companies to provide value. 3. As a user, you should be aware of the solution's limitations and your requirements. 4. If you are a company, you should also be aware of the software and hardware lock-in of the commercial solutions. Before choosing your simulation stack think deeply about these questions: 1. Why do I need a simulator? 2. What are the most important physics phenomena that I want to simulate? 3. What is the scale of use for my simulator? 4. What are my requirements for software complexity and integrations? Don't forget Occam's razor while answering these questions :)

Project GR00T is just the tip of the iceberg! NVIDIA's Isaac platform just got a MASSIVE upgrade, making robot development faster, easier and more powerful than ever before. Here's the TL;DR: GR00T: Train robots with just a few demonstrations! Say goodbye to complex programming. Isaac Lab: Run thousands of simulations in parallel for lightning-fast robot learning. ⚡️ OSMO: Scale your robot development across any environment with ease. Isaac Manipulator: Unlock next-level dexterity and AI capabilities for robotic arms. (Up to 80x speedup in path planning!) Isaac Perceptor: Give your robots 3D surround vision for safer, more efficient operations. This is a game-changer for robotics! The future is here, and it's built on NVIDIA. #NVIDIA #GR00T #Isaac #Robotics #TheFutureIsNow

Our team recently dove into Nvidia’s Omniverse and Isaac Sim to see how these platforms are shaping the future of robotics. Here’s what we found: Omniverse is carving out its niche in realistic scene and object modeling, rivaling giants like Unity and Unreal Engine. Its extensions for motion simulation, like PhysX and RTX Sensors, make it a powerful tool. However, the transition from simulation to real-world robotics isn't seamless. The communication methods between simulators and real sensors/motors often differ, posing a challenge. Gazebo, integrated with ROS, offers a smoother experience in this regard. Isaac Sim, on the other hand, is becoming a go-to for generating synthetic data, especially for LCBM (Large Content and Behavior Models) solutions. While it’s excellent for this purpose, it requires specific Nvidia-compatible hardware and lacks the abstraction level needed for broader compatibility. This makes it tough to transfer learnings across different hardware components. In our experiments, we’ve seen the effectiveness of synthetic data firsthand, training a traffic sign classifier with it and winning multiple awards in the AI GO competitions using Gym-Duckietown. Isaac Sim is promising, but understanding where to use it versus testing on hardware is key to leveraging its full potential. #Robotics #SyntheticData #Simulation #AutonomousSystems

Nvidia made some huge robotics announcements at their GTC 2024 conference. The headliner was GR00T, a novel foundation model designed to power the next wave of advanced humanoid robots. The goal is for GR00T to allow humanoid robots to understand natural language, mimic human movements with dexterity, and operate intelligently in the real world. However, many details around GR00T's inner workings and training data remain vague. To run GR00T and other robotics AI workloads, Nvidia also unveiled their new Jetson Thor computing platform featuring their latest Blackwell GPU architecture. #GR00T #HumanoidRobots #ArtificialGeneralIntelligence #RobotOperatingSystem #ROS #OpenSource #OpenRobotics #JetsonThor #BlackwellGPU #Robotics #AI #FoundationModels #OSRA #NVIDIA #GTC2024 #FutureOfRobotics

Gazebo or Isaac Sim The robotics simulation tools are undergoing a massive shift with the latest push from NVIDIA towards building a community around Isaac sim and many companies migrating to it. The choice remains unclear for many novice users. Here are some things I’ve learned working with both: * Isaac sim offers significantly better photo realism, which is key for data collection and model training. * Isaac sim offers massive parallelization capabilities over multiple GPUs, which makes it advantageous for RL pipelines, where simulation scale is important. * Physics of Isaac sim are still lacking for modeling certain types of contacts in a stable manner, but it’s under active development. * Gazebo has been a traditional choice due to its straight forward ROS integration. However, the latest versions of Isaac sim offer bridges to ROS2. I expect this interface to improve as many robotics companies are actively using it. In conclusion, Gazebo has had a good run and been the default choice for many years. But, Isaac sim is definitely gaining momentum with a fast growing community. Would love to hear other perspectives from folks who have used either or both. #isaacsim #robotics #simulation #gazebo #ros #roboticsengineering

#GTC2024 😎 moments review: Join Roboflow and explore how to seamlessly deploy YOLO-World zero-shot object detection model on #reComputer Industrial edge device powered by NVIDIA Robotics Jetson Orin Nano! The whole pipeline transcends traditional boundaries with its open vocabulary capability, allowing it to recognize objects beyond predefined categories—making it more efficient and adaptable for real-world applications. Experience the flexibility of dynamically adjusting detection vocabularies to meet diverse needs without compromising performance. 👉 To get hands on the YOLO-World approach on Jetson Orin Nano 8GB, you may want to check out our reComputer Industrial J3011: https://lnkd.in/gcfq6g8D 🏂 YOLO-World GitHub repo: https://lnkd.in/gvASzQUX #nvidia #jetson #computervision #edgeai #zeroshot #objectdetection #yoloworld

For developers working on #robotics projects with vision processing, integrating a Pi camera with #ROS2 Humble on a Jetson Nano seamlessly could be a fundamental step, in preparing the ready-to-use kit for further robotics development. Here's the step-by-step guidance from Kabilan Kb about how to connect the camera with NVIDIA Robotics Jetson Nano developer kit, and how to build ROS nodes for image processing in robotics applications 👉 https://lnkd.in/g7APpTt8 Additionally, discover the compatible CSI camera list for Jetson devices (Orin NX/Orin Nano/Nano): https://lnkd.in/gc6ftDpm

🤖 Nvidia's Game-Changer: Project GR00T Ushers in the Era of Humanoid Robots! 🚀 At the heart of innovation, Nvidia made headlines at the GTC Developer Conference with the electrifying launch of Project GR00T, a groundbreaking initiative that propels the tech titan further into the forefront of AI and robotics. This bold move marks Nvidia's entry into the realm of humanoid robots, positioning it as a catalyst for the next wave of technological evolution. Project GR00T is more than just a platform; it's a visionary ecosystem designed to supercharge the capabilities of robot makers. Partners like Agility Robotics and Sanctuary AI are already lined up to harness the power of Nvidia's latest silicon masterpiece, Jetson Thor, alongside cutting-edge programs such as Isaac Manipulator and Isaac Perceptor. These innovations are set to redefine the landscape of robotics, empowering humanoid robots with unprecedented levels of intelligence and versatility. But what's truly exciting is the potential showdown Nvidia is setting the stage for: humanoid robots versus mobile manipulators. This isn't just about technological prowess; it's about shaping the future of human-robot interaction and the role robotics will play in our everyday lives. 🌟 Why is this significant? Nvidia's foray into humanoid robotics with Project GR00T signals a major shift in how we envision the integration of robots in society. It's a testament to Nvidia's commitment to pushing the boundaries of what's possible in AI and robotics, heralding a new era of innovation and collaboration. #Nvidia #ProjectGR00T #HumanoidRobots #AI #RoboticsInnovation #FutureOfTech Let's dive into the discussion: How do you see humanoid robots impacting our future? What opportunities and challenges do you envision as we step into this new era of robotics? Share your thoughts and insights below! https://lnkd.in/eWnFz6zV