{"id":23865,"date":"2018-04-24T11:03:44","date_gmt":"2018-04-24T15:03:44","guid":{"rendered":"https:\/\/www.tun.com\/blog\/?p=23865"},"modified":"2022-03-16T12:02:11","modified_gmt":"2022-03-16T16:02:11","slug":"computer-animation-algorithm-breakdancing-acrobatic-simulated-characters","status":"publish","type":"post","link":"https:\/\/www.tun.com\/blog\/computer-animation-algorithm-breakdancing-acrobatic-simulated-characters\/","title":{"rendered":"New Algorithm Leads Tt Breakdancing, Acrobatic Simulated Characters"},"content":{"rendered":"

A team of researchers from the University of California, Berkeley and the University of British Columbia in Canada has <\/span>developed an algorithm to re-create natural motions in computer animation<\/span><\/a>. <\/span><\/p>\n

Traditional computer-simulated motions are seen as clumsy and rhythmless, often failing at mimicking a human\u2019s natural motions. <\/span><\/p>\n

Disappointed by old techniques, the team was inspired to find a solution. <\/span><\/p>\n

\u201cThe motivation for this work is that we want to develop simulated characters that can perform some very challenging skills while also moving in a natural manner,\u201d said <\/span>Xue Bin (Jason) Peng<\/span><\/a>, a UC Berkeley graduate student and researcher. <\/span><\/p>\n

Pieter Abbeel<\/span><\/a> and <\/span>Sergey Levine<\/span><\/a>, both from the UC Berkeley Department of Electrical Engineering and Computer Sciences, and <\/span>Michiel van de Panne<\/span><\/a> from the University of British Columbia also contributed to the study.<\/span><\/p>\n

The researchers used deep reinforcement learning to re-create natural motions in humans. With this technique, the simulated characters can do acrobatics, breakdancing and martial arts, and can even respond to changes in the environment, such as being tripped or dodging projectiles.<\/span><\/p>\n

The Computer System (DeepMimic)<\/b><\/h2>\n

Traditionally, there have been two techniques used in computer animation. <\/span><\/p>\n

One requires designing customized controllers for every skill, such as walking, flipping, or running. The results from this method usually look pretty good, Peng said. <\/span><\/p>\n

The other technique, which uses deep reinforcement learning methods, can simulate many tricks using a single algorithm, but its results often look unnatural. <\/span><\/p>\n

The researchers\u2019 new technique allows them to get \u201cthe best of both worlds,\u201d Peng said in a statement. <\/span><\/p>\n

The team\u2019s algorithm can simulate many tricks, and could surpass the appearance of traditional hand-controller methods. <\/span><\/p>\n

\u201cOur method is extremely simple,\u201d said Peng.<\/span><\/p>\n

\u201cWe first collect a single demonstration of a skills from a human (ex. backflip or spin kick),\u201d he continued. \u201cThe demonstrations are usually in the form of motion capture clips. We then feed this demonstration to a reinforcement learning algorithm that tries to imitate the motion of the human. The agent imitates the motion by minimizing the tracking error at every timestep, and this simple approach ends up allowing the character to learn some very dynamic and acrobatic skills.\u201d<\/span><\/p>\n

Peng collected reference data from more than 25 motion-capture clips of backflips, cartwheels, kip-ups, vaults, running, throwing, jumping, and more. <\/span><\/p>\n

The team then allowed the system, named DeepMimic, to practice each skill for around a month. <\/span><\/p>\n

The computer practiced all day and night and used trial and error to find the closest match to real human movements. <\/span><\/p>\n

Because difficult movements, such as the human backflip, require many individual body movements, the researchers set the algorithm to learn various stages of the backflip. It then took all of the stages and stitched them together to create a full motion. <\/span><\/p>\n