Towards the robots of science fiction | Aaron Ames | TEDxManhattanBeach

In this impressive and entertaining talk, Aaron Ames shares the challenge and complexity of the mathematics and physics behind human-like, bipedal, robotic walking. He also introduces Cassie, a bipedal robot, who demonstrates her walking on stage. Aaron Ames runs the Advanced Mobility Laboratory (AMBER Lab) as part of the Department of Mechanical and Civil Engineering, the Department of Computing and Mathematical Science and at Caltech’s Center for Autonomous Systems and Technologies (CAST). In the AMBER Lab, he and his students hand-build and test bipedal robots and prosthetic limbs, and create the theory and the algorithms that govern how the robots walk. The goal is to achieve human-like robotic walking and translate that capability to robotic assistive devices as well as robots that can explore environments not accessible to humans-including Mars and beyond.
Ames earned his undergraduate degrees from the University of St. Thomas in St. Paul, Minnesota, and his MA and PhD from the University of California, Berkeley. In 2010, he received the National Science Foundation CAREER award for his research on bipedal robotic walking and its applications to prosthetic devices. He also received the 2015 Donald P. Eckman Award. This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at www.ted.com/tedx

[@indranilbiswas629]

Caltech rocks ❤️

[@engineering_chaos2749]

Amazing. If only the algorithm was publicly available, I would attempt to make a Cassie of my own.

[@AmericanStuff2024]

Engineering: Perhaps in your lifetime.

[@regularguy8110]

Boston Dynamics learning curve is a toddler learning curve. After raising 6 kids...It looks exactly like kids learning to control their body. The back flip fail happens thousands of time at kids gymnastics classes. To me, the difference is when AI catches up with the learning curve. It uses logic not muscle memory and once learned is precisely repeatable and essentially permanent. It won't have a relearning/refresh curve once it's coded.

[@hassansaeed3181]

Wow !

[@joannot6706]

It's must have taken a lot of efforts to get to their results and this is a great achievement.
Nonetheless I have seen way more dynamic algorithms applied on Cassie's hardware:
Namely using reinforcement learning rather than the approach they seem to use that looks quite labor intensive as opposed to programming an AI that will figure out how to walk in a simulated environment without necessarily teaching specific rules about walking.
Maybe there is something I misunderstand about their methods but it seems like machine learning is easier given the proper compute.

[@Jorfy]

Yea I agree the walking could be better

[@ethanbarlow1804]

Man, what I'd give to work in a research lab like that . . .

[@AmericanStuff2024]

Ethan: Apply to graduate school in robotics. Takes very good math, calculus.

[@ethanbarlow1804]

@@AmericanStuff2024 I've applied to several hoping to start this fall.🤞

[@Martin-jp8pw]

@@ethanbarlow1804 Did you start?

[@factandsciencetracker5860]

Search on Amazon : The Unseen Realms Of The Universe

[@KissMyFatAxe]

So basically we're gonna become cyborgs. Count me in 🙌
In all seriousness though, this is revolutionary stuff. It can improve quality of life for millions of people worldwide. This is what human innovation is all about. We're smarter than we give ourselves credit for
Welcome to the future! 🤣🙌

[@AmericanStuff2024]

Born: People are now counting how many robotic parts they have. I have one piece. One of my brothers has 2.

[@LuisCastillo-bl6mq]

All was interesting but he forgot that the robot that shows at first, was just made by a computer, i mean it is not real, there is a video in youtube which shows you the video behind scene

[@doomguy2532]

Look like metal gear

[@AbuChanOfficial]

Science fiction won't be fiction

[@solangerego992]

Dude, this is just fascinating, exciting and QUITE SCARRY. I am waiting for the day the Japanese will put a machine gun on those robots.

[@MikenNinginThai]

That is the danger of Ai as you see with drones by the US using Ai in a bad way not really helping humanity.

[@Tayo39]

tellem to say hello to my drone army of 500 ranging from the size of a mosquito to a fighterjet, locked, loaded, connected, hovering around & above me while looking for dangers autonomously, 24/7/365

[@AmericanStuff2024]

Solangerego: USA already has a state of the art robot army but it flies: drones.

[@AmericanStuff2024]

@@MikenNinginThai Depends completely on a person's perspective. USA will not now have a fascist civil war nor a fascist revolution because the USA Military has been given a state of the art drone army.
Now, what the capacity becomes in the future is a function of what the voters do and what the USA Military does. The USA Military oath of service is to defend and protect the Constitution. In doing so, it is legal for them to disobey presidential commands.
IF the voters were to empower a fascist administration, the USA Military has the capacity to resist it.
What has happened is that the USA now has 5 branches of government rather than 3:
1) Voters
2) Legislature
3). Executive/Civil Service
4) Supreme Court
5) Military
The military would win because it controls the actual, physical weapons

[@bpk8320]

cannot let the chin return to china !