RF Grasp: The robot that finds its way through clutter
Robots are not capable of handling tasks as simple as restocking grocery store shelves as they cannot perceive the environment as good as humans. What if we could give robots radio perception to search for items that are not in their sight? Such an ability will give them superhuman power to work in warehouses, stores, and our homes.
The Signal Kinetics research group present the design, implementation, and evaluation of RF-Grasp, a robotic system that can grasp fully occluded objects in unknown and unstructured environments. Unlike prior systems that are constrained by the line-of-sight perception of vision and infrared sensors, RF-Grasp employs RF (Radio Frequency) perception to identify and locate target objects throughocclusions, and perform efficient exploration and complex manipulation tasks in non-line-of-sight settings.
RF-Grasp relies on an eye-in-hand camera and batteryless RFID tags attached to objects of interest. It introduces two main innovations: (1) an RF-visual servoing controller that uses the RFID’s location to selectively explore the environment and plan an efficient trajectory toward an occluded target, and (2) an RF-visual deep reinforcement learning network that can learn and execute efficient, complex policies for decluttering and grasping.
The research team implemented and evaluated an end-to-end physical prototype of RF-Grasp and a state-of-the-art baseline. They demonstrated it improves success rate and efficiency by up to 40-50% in cluttered settings, and also demonstrate RF-Grasp in novel tasks such mechanical search of fully occluded objects behind obstacles, opening up new possibilities for robotic manipulation.
Learn more: https://www.media.mit.edu/projects/rfgrasp/overview/
Видео RF Grasp: The robot that finds its way through clutter канала MIT Media Lab
The Signal Kinetics research group present the design, implementation, and evaluation of RF-Grasp, a robotic system that can grasp fully occluded objects in unknown and unstructured environments. Unlike prior systems that are constrained by the line-of-sight perception of vision and infrared sensors, RF-Grasp employs RF (Radio Frequency) perception to identify and locate target objects throughocclusions, and perform efficient exploration and complex manipulation tasks in non-line-of-sight settings.
RF-Grasp relies on an eye-in-hand camera and batteryless RFID tags attached to objects of interest. It introduces two main innovations: (1) an RF-visual servoing controller that uses the RFID’s location to selectively explore the environment and plan an efficient trajectory toward an occluded target, and (2) an RF-visual deep reinforcement learning network that can learn and execute efficient, complex policies for decluttering and grasping.
The research team implemented and evaluated an end-to-end physical prototype of RF-Grasp and a state-of-the-art baseline. They demonstrated it improves success rate and efficiency by up to 40-50% in cluttered settings, and also demonstrate RF-Grasp in novel tasks such mechanical search of fully occluded objects behind obstacles, opening up new possibilities for robotic manipulation.
Learn more: https://www.media.mit.edu/projects/rfgrasp/overview/
Видео RF Grasp: The robot that finds its way through clutter канала MIT Media Lab
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Meet Spot, the robot dog that can run, hop and open doors | Marc RaibertProgrammable DropletsNew bionics let us run, climb and dance | Hugh HerrkinetiX—designing auxetic-inspired deformable material structuresOptimizing your Brain through ExerciseSeeding Innovation: The MIT Media Lab at Aspen Ideas Festival100 robots "Pepper" support the SoftBank Hawks baseball team with cheer danceSample efficient reinforcement learning with SAC on Kuka robotML30: Penn & Teller TrickINSANE Domino Tricks! (Hevesh5 & MillionenDollarBoy)Amazing Technology Invented By MIT - Tangible MediaAlterEgo: Interfacing with devices through silent speech9 MIT Media Lab Innovations that Changed the FutureTransformative AppetiteRobots of the future at Boston DynamicsHow to Start a SpeechReusable support for 3D printersMAB Robotics - first steps of the Bare Metal Walking Robot PrototypeVision-free MIT CheetahHow MIT Builds Cities Using Lego and Augmented Reality | Science of Teams | WIRED