Highly Dynamic Quadruped Locomotion via Whole-Body Impulse Control and Model Predictive Control

Donghyun Kim, Jared Di Carlo, Benjamin Katz, Gerardo Bledt, and Sangbae Kim
Highly Dynamic Quadruped Locomotion via Whole-Body Impulse Control and Model Predictive Control
Paper available at arxiv.org/abs/1909.06586

[@heinrichwonders8861]

Impressive for such a compact design.
Are you planning to add dynamic gait-switching?

[@MrFox_Adventures]

Fantastic project!

[@Robotib]

😃 massa demais !

[@georgehancock652]

From a biomimetic perspective why two joint rather than 3 joint limbs, as currently the stride length is really short. I would’ve thought that having a three joint limb would increase the stride engraved and well the amount of spring. If it’s truly beneficial to have a two joint. Limb then why do we only see knee driven locomotion in birds for which all modern ground dwelling species are secondarily terrestrial?

[@literallymothman4596]

Maybe it was too heavy then? Idk I don’t know much about robots or anything but maybe they wanted it to be small, but fast? Again, not a robot person.

[@ethanchandler2792]

Each leg has 3 joints, not 2. It would not be able to turn as dynamically without the abduction joint. Read the paper

[@raguaviva]

Cheetah 2 from 2016 was achieving 5.8m/s, this one only 3.7m/s

[@liuhorson6207]

They are in different sizes

[@germanblengini7577]

Why was the leg movement design changed from the previous version ?! is it more efficient or just because it is more adaptable?

[@dougg1075]

It will all be over soon BWAHAHAHAA!