Technical Directions
Robotics coverage relating to robot policies.
335 selected reportsLast 30 days 289 recordsAll coverage 1,886 records
The US is banning the sale of foreign-made robots, including those from China, due to national security concerns. The FCC cites risks such as vulnerabilities in networked capabilities and potential exploitation of remote connectivity. The ban also extends to inverters with Wi-Fi, cellular, or Bluetooth control. Existing products already sold are unaffected, but companies can apply for approval to continue selling in the US.
LeRobot has converted the Hy-Embodied-0.5-VLA-RoboTwin model, achieving full forward parity with the original policy. The model supports 16D dual-arm pose/gripper actions and passes all validation checks, though it cannot handle more complex action spaces.
LeRobot has released a converted version of the Hy-Embodied-0.5-VLA-UMI model, ensuring full tensor compatibility and validation. The model supports a 16D dual-arm embodiment but cannot express 12D mobile-manipulator actions, indicating limitations in the current representation.
Google DeepMind has released Gemini Robotics 2, a new model that enables robots to perform complex tasks with whole-body intelligence, advanced dexterity, and multi-robot collaboration. The model supports real-time reasoning, on-device adaptation, and safety features for human-robot interaction.
NVIDIA introduces Cosmos-H-Dreams, a real-time, action-conditioned generative simulator for surgical robotics. It distills the capabilities of Cosmos-H-Surgical-Simulator into a causal, few-step student model, enabling interactive environments for policy evaluation and synthetic data generation. The system runs on a single NVIDIA RTX PRO 6000 GPU and supports integration with surgical platforms like Versius.
NVIDIA has released an open-source, GPU-accelerated Medical Physics Simulation framework as part of its Isaac for Healthcare platform. This tool enables developers to model anatomy-device interactions, generate complex scenarios, and train robot policies in simulation, significantly reducing development time and improving regulatory readiness.
Engineers at MIT and EPFL have developed a flapping-wing robot, or FAAV, that can swim underwater and then flap into the air, mimicking the movement of diving birds. The robot, weighing less than 300 grams, can transition between water and air using optimized wing size, flapping frequency, and tail angle. The study, published in Science, could enable new aerial-aquatic drones for oceanographic research and environmental monitoring.
NVIDIA and Hugging Face are releasing the NVIDIA Isaac GR00T 1.7 model and Isaac Teleop framework into LeRobot, an open-source robotics library, to provide developers with shared tools for training, evaluating, and deploying robot foundation models. NVIDIA Cosmos 3, a frontier world model for physical AI, is also set to be integrated soon.
NVIDIA has released a cross-embodiment foundation model for humanoid robots, leveraging a Cosmos-Reason2/Qwen3-VL backbone and a flow-matching action transformer. The model is trained using LeRobot and is available for training and deployment under the Apache 2.0 license.
NVIDIA has released a cross-embodiment foundation model for humanoid robots, leveraging a Cosmos-Reason2/Qwen3-VL backbone and a flow-matching action transformer. The model is trained using LeRobot and is available for training and deployment via Hugging Face.
NVIDIA has released the 'gr00t17-lerobot-libero_object-640' model, a cross-embodiment foundation model for humanoid robot reasoning and skills. It uses a Cosmos-Reason2/Qwen3-VL backbone and a flow-matching action transformer, trained with LeRobot and available under the Apache 2.0 license. The model supports training and deployment through LeRobot, with detailed instructions provided.
NVIDIA has released a cross-embodiment foundation model for humanoid robots, leveraging a Cosmos-Reason2/Qwen3-VL backbone and a flow-matching action transformer. The model is trained using LeRobot and is available for training and deployment via the Hugging Face Hub.
Boston Dynamics explores how football can inform robot movement by using human demonstrations and reinforcement learning to train Atlas to perform a complex 'Ghost Rabona' kick, highlighting the intersection of sports and robotics in developing dynamic, whole-body control.
Skild AI has acquired Zebra Technologies' robotics division to deploy its omni-bodied brain across warehouses, aiming to unlock massive productivity gains and enhance end-to-end fulfillment processes. This acquisition integrates Zebra's Symmetry Fulfillment platform with Skild's foundation model, enabling cross-embodiment generalization and intelligent decision-making in logistics environments.
Dexterity has deployed Instinct, a tactile intelligence system that enables robots to perform force-guided manipulation with zero retraining. Instinct reacts to contact forces in 0.5–2 ms, 10–100× faster than human touch, and adapts in real time to constrained placements. It integrates with Dexterity’s world model, Foresight, to update contextual understanding through contact data, achieving an 81% success rate in production with adaptive strategies.
Google DeepMind has released Gemini Robotics-ER 1.6, an upgraded model that enhances embodied reasoning for robotics. This model improves spatial reasoning, multi-view understanding, and safety compliance, enabling robots to perform complex tasks like instrument reading. It is available via the Gemini API and Google AI Studio, with examples provided in a developer Colab.
Helix 02, a single neural system that controls the full body directly from pixels, now performs end-to-end living room cleanup, including tasks like cleaning surfaces, handling flexible objects, and bimanual manipulation. The system learns new capabilities through data without requiring new algorithms or specialized engineering.
This tutorial explores the challenges of deploying VLA models on embedded robotic systems, including dataset recording best practices, fine-tuning techniques for ACT and SmolVLA, and real-time performance optimization using the NXP i.MX 95 SoC. It emphasizes asynchronous inference and hardware-aware scheduling to improve control and reduce latency.
Dexterity introduces Foresight, a world model that enables robots to reason about the physical world, predict outcomes, and act confidently in real environments. Trained on over 100 million autonomous actions in production, Foresight is designed for interpretability, safety, and performance, with capabilities like predictive branching and pragmatic decision-making. It addresses complex tasks like spatial packing and includes a game and API challenge for further exploration.
Physical AI refers to AI systems that enable robots to perform physical tasks in the real world. The article explains that this is extremely challenging due to the infinite variability of the physical environment, requiring solutions across perception, planning, motion control, and safety. Dexterity's compositional approach, using specialized AI agents and an orchestrator, enables production-scale Physical AI with over 100 million autonomous actions executed in real enterprise operations.
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).