A fun paper expanding on the control methods used in the well-known and well-performing Atlas robots by Boston Dynamics. A little bit of a ‘behind the curtain’ result, it’s interesting to see the methods behind a robot that has gained so much attention. Unsprisingly, online optimization of trajectory outcomes is employed, suggesting that a more constructive theory for robot decision making and control is missing.
Link to Paper
Authors : Scott Kuindersma, Robin Deits, Maurice Fallon, Andrés Valenzuela, Hongkai Dai, Frank Permenter, Twan Koolen, Pat Marion, and Russ Tedrake
Journal : Under Review.
This paper describes a collection of optimization algorithms for achieving dynamic planning, control, and state estimation for a bipedal robot designed to operate reliably in complex environments. To make challenging locomotion tasks tractable, we describe several novel applications of convex, mixed-integer, and sparse nonlinear optimization to problems ranging from footstep placement to whole-body planning and control. We also present a state estimator formulation that, when combined with our walking controller,permits highly precise execution of extended walking plans over non-flat terrain.We describe our complete system integration and experiments carried outon Atlas, a full-size hydraulic humanoid robot designed by Boston Dynamics.