Method of Controlling a Robot

摘要

A method of controlling motion on the ground of a device having at least two legs comprising: obtaining current values of dynamic state variables of the device; optimizing a movement command for the device under constraints while taking account of said current values and of a movement setpoint for the device, said optimization being performed by modeling a movement of the device between a current state and a second state, the device being supported during said movement by at least on a first leg, and then from an instant in which the center of pressure on the ground is changed, at least on a second leg; and applying said movement command; wherein said optimization is performed on the assumption that the gravity potential energy of the device in the second state is a local maximum as a function of time.

主权项

1. A method of controlling motion on the ground of a device having at least two legs comprising:
obtaining current values of dynamic state variables of the device; an optimization of performing a movement command for the device under constraints based at least partially on said current values and a movement setpoint for the device, said optimization being performed by modeling a movement of the device between a current state and a second state, the device being supported during said movement by at least a first leg, and during an instant in which a center of pressure of the device on the ground is changed, the device being supported by at least a second leg; and applying said movement command; wherein said optimization is performed based at least partially on an assumption that the gravitational potential energy of the device in the second state is a local maximum as a function of time.