Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Using reinforcement learning we have expanded the range of techniques the Ultra Mobile Vehicle (UMV) uses to handle terrain and obstacles, including hops, out-of-plane balance, and level-ground flips. Millions of physics-based simulations provide training data to support zero-shot transfers.

rai_inst's profile picture

RAI Institute

8,419 subscribers

75,272 Aufrufe • vor 1 Jahr •via X (Twitter)

Wissenschaft & TechnologieGamingBildung
LIVE

Anya Rossi• Live Now

Private livecam show

17 Kommentare

Profilbild von RAI Institute
RAI Institutevor 1 Jahr

Learn more about how we’re building and training UMV:

Profilbild von GrayWolf448
GrayWolf448vor 1 Jahr

Can you guys build the torment nexus next?

Profilbild von Master Kenobi
Master Kenobivor 1 Jahr

Hey @danny_macaskill 😉

Profilbild von Dan Calle
Dan Callevor 1 Jahr

Oof. Ok. Amazing tech, and all I can think of are @itsDanielSuarez's razorbacks.

Profilbild von Harshith
Harshithvor 1 Jahr

@steveMmattison Wtf is this...

Profilbild von Crypto
Cryptovor 1 Jahr

in robots ko car ko drift karne se rokna hai bass..arrre tu robot hai sidhe bhaag bhai..cycle kyun chala rha hai.

Profilbild von NINJA RANGER
NINJA RANGERvor 1 Jahr

so where can you mount the gatling gun on?

Profilbild von Boris
Borisvor 1 Jahr

Terminator...

Profilbild von Chi Tinh Mai
Chi Tinh Maivor 1 Jahr

That UMV sounds like it’s ready for the next season of “Extreme Terrain Challenge”!

Profilbild von Monocle Optical Labs . . 👊🏾`.')
Monocle Optical Labs . . 👊🏾`.')vor 10 Monaten

RAD ✌🏾😎

Profilbild von Raul Verdusco
Raul Verduscovor 1 Jahr

Reinforcement learning continues to unlock incredible capabilities for mobile robotics! ⚡🤖 Training UMVs through millions of physics-based simulations enables zero-shot transfers — giving robots advanced techniques like hops, flips, and out-of-plane balance for real-world terrain. 🌍🦾 A strong step forward in creating more adaptable, resilient, and versatile robotic platforms. 🚀🔧 #Robotics #ReinforcementLearning #AI #Automation #Engineering #Innovation #SmartTechnology #FutureOfWork

Profilbild von iRiSh
iRiShvor 1 Jahr

What are you doing ? Who needs this ? I'm sorry but we all learned from Arc Raiders that the answer is an A.I trained giant 6 legged spider-crab with lasers for a mouth ... stop wasting time. And Yes ... I am available to hire for more great ideas like this.

Profilbild von euno
eunovor 1 Jahr

Most people cannot do that, neither can they ride on one wheel.

Profilbild von Rough
Roughvor 1 Jahr

bonkers

Profilbild von advocAIdo
advocAIdovor 1 Jahr

Just add training wheels, but I guess that will kill the cool factor.

Profilbild von arie moby
arie mobyvor 1 Jahr

You also show the failure. Such honesty is not in China.

Profilbild von Roger Goll
Roger Gollvor 1 Jahr

But seriously: Great project! Huge military and commercial potential. The market will change significantly in the coming years.

Ähnliche Videos

Reinforcement learning is used to speed the production of behavior for the Boston Dynamics Atlas humanoid robot. At the heart of the learning process is a physics-based simulator that generates training data for a variety of maneuvers.
1:10

Reinforcement learning is used to speed the production of behavior for the Boston Dynamics Atlas humanoid robot. At the heart of the learning process is a physics-based simulator that generates training data for a variety of maneuvers.

RAI Institute

76,796 Aufrufe • vor 1 Jahr

Soccer players have to master a range of dynamic skills, from turning and kicking to chasing a ball. How could robots do the same? ⚽ We trained our AI agents to demonstrate a range of agile behaviors using reinforcement learning. Here’s how. 🧵
0:27

Soccer players have to master a range of dynamic skills, from turning and kicking to chasing a ball. How could robots do the same? ⚽ We trained our AI agents to demonstrate a range of agile behaviors using reinforcement learning. Here’s how. 🧵

Google DeepMind

448,277 Aufrufe • vor 2 Jahren

Φ-SO : Physical Symbolic Optimization - Learning Physics from Data 🧠 The Physical Symbolic Optimization package uses deep reinforcement learning to discover physical laws from data. Here is Φ-SO discovering the analytical expression of a damped harmonic oscillator.
0:20

Φ-SO : Physical Symbolic Optimization - Learning Physics from Data 🧠 The Physical Symbolic Optimization package uses deep reinforcement learning to discover physical laws from data. Here is Φ-SO discovering the analytical expression of a damped harmonic oscillator.

Jousef Murad

433,257 Aufrufe • vor 3 Jahren

RAI Institute’s Ultra mobility vehicle is showing off advanced new stunts with smooth 360-degree spins, kip jumps, flips, and bunny hops. The AI-powered machine performs sharp turns, mid-air rotations, and clean landings with impressive balance and precision. And this is only the beginning of what the vehicle can achieve.
0:25

RAI Institute’s Ultra mobility vehicle is showing off advanced new stunts with smooth 360-degree spins, kip jumps, flips, and bunny hops. The AI-powered machine performs sharp turns, mid-air rotations, and clean landings with impressive balance and precision. And this is only the beginning of what the vehicle can achieve.

Space and Technology

57,164 Aufrufe • vor 5 Monaten

Robots can now walk blind with Figure AI's Helix Walking Controller. Instead of using cameras, the controller relies on the robot's internal sensors and reinforcement learning to maintain balance while walking. The controller, unveiled last year, allows the robot to stay balanced, recover from pushes, and move confidently across difficult terrain.
0:26

Robots can now walk blind with Figure AI's Helix Walking Controller. Instead of using cameras, the controller relies on the robot's internal sensors and reinforcement learning to maintain balance while walking. The controller, unveiled last year, allows the robot to stay balanced, recover from pushes, and move confidently across difficult terrain.

Space and Technology

28,912 Aufrufe • vor 4 Monaten

Visuals from the ground: Indian military personnel help in trying to provide support in the aftermath of the Ahmedabad plane crash.
0:25

Visuals from the ground: Indian military personnel help in trying to provide support in the aftermath of the Ahmedabad plane crash.

Sidhant Sibal

196,760 Aufrufe • vor 1 Jahr

Robot policies must be both reliable and highly capable to be useful; the best way to achieve this level of performance is with reinforcement learning. However, for reinforcement learning you are usually stuck between two difficult options: reinforcement in the real world is often risky and expensive, while reinforcement learning in a traditional simulator takes a lot of engineering work and has a persistent sim-to-real gap. What if instead you could train your robot purely in a world model? RISE by Jiazhi Yang et al. uses a compositional world model to predict the future and evaluate progress. This allows for a self-improving pipeline, which learns a world model from real data and then learns how the robot should perform different tasks. This pipeline results in a data-driven way to improve policy performance from real data but without real-world reinforcement learning. Watch Episode #86 of RoboPapers, with Chris Paxton and Jiafei Duan, to learn more!
53:45

Robot policies must be both reliable and highly capable to be useful; the best way to achieve this level of performance is with reinforcement learning. However, for reinforcement learning you are usually stuck between two difficult options: reinforcement in the real world is often risky and expensive, while reinforcement learning in a traditional simulator takes a lot of engineering work and has a persistent sim-to-real gap. What if instead you could train your robot purely in a world model? RISE by Jiazhi Yang et al. uses a compositional world model to predict the future and evaluate progress. This allows for a self-improving pipeline, which learns a world model from real data and then learns how the robot should perform different tasks. This pipeline results in a data-driven way to improve policy performance from real data but without real-world reinforcement learning. Watch Episode #86 of RoboPapers, with Chris Paxton and Jiafei Duan, to learn more!

RoboPapers

38,867 Aufrufe • vor 5 Monaten

RAI’s Ultra Mobility Vehicle uses AI, sensors, and dynamic motion to stay stable and tackle extreme terrain.
1:01

RAI’s Ultra Mobility Vehicle uses AI, sensors, and dynamic motion to stay stable and tackle extreme terrain.

Science girl

12,467 Aufrufe • vor 7 Monaten

Footage of a Ukrainian soldier using a Steam Deck to control the weapons system of an unmanned ground vehicle at a firing range.
0:26

Footage of a Ukrainian soldier using a Steam Deck to control the weapons system of an unmanned ground vehicle at a firing range.

OSINTtechnical

75,274 Aufrufe • vor 9 Monaten

Boston Dynamics Atlas robot recently performed flips and cartwheels after extensive training and testing. Engineers used simulations and repeated practice runs to improve its balance, timing, and coordination. By learning from mistakes and refining its movements, Atlas became stable enough to perform the actions smoothly.
0:33

Boston Dynamics Atlas robot recently performed flips and cartwheels after extensive training and testing. Engineers used simulations and repeated practice runs to improve its balance, timing, and coordination. By learning from mistakes and refining its movements, Atlas became stable enough to perform the actions smoothly.

Space and Technology

83,781 Aufrufe • vor 5 Monaten

Zero-shot navigation of complex terrain. The true advantages of humanoid robots are their ability to work in the true difficulty of human environments. All using onboard sensing and compute. Take a look ->
0:46

Zero-shot navigation of complex terrain. The true advantages of humanoid robots are their ability to work in the true difficulty of human environments. All using onboard sensing and compute. Take a look ->

Chris Paxton

10,945 Aufrufe • vor 1 Monat

Simulating entry… Using our LAVA software, Ames engineers performed simulations of a conceptual Mars spacecraft during entry. Here, the vehicle travels at supersonic speed (Mach 2.4) and uses retro-propulsion to decelerate further.
1:03

Simulating entry… Using our LAVA software, Ames engineers performed simulations of a conceptual Mars spacecraft during entry. Here, the vehicle travels at supersonic speed (Mach 2.4) and uses retro-propulsion to decelerate further.

NASA Ames

80,589 Aufrufe • vor 3 Monaten

Helix is now learning directly from human video data We have already trained on data collected in the real world, including Brookfield residential units To our knowledge, this is the first instance of a humanoid robot learning navigation end-to-end using only human video
0:24

Helix is now learning directly from human video data We have already trained on data collected in the real world, including Brookfield residential units To our knowledge, this is the first instance of a humanoid robot learning navigation end-to-end using only human video

Figure

46,041 Aufrufe • vor 1 Jahr