TY - GEN
T1 - Full dynamics LQR control of a humanoid robot
T2 - 2014 14th IEEE-RAS International Conference on Humanoid Robots, Humanoids 2014
AU - Mason, Sean
AU - Righetti, Ludovic
AU - Schaal, Stefan
N1 - Publisher Copyright:
© 2014 IEEE.
PY - 2015/2/12
Y1 - 2015/2/12
N2 - Humanoid robots operating in human environments require whole-body controllers that can offer precise tracking and well-defined disturbance rejection behavior. In this contribution, we propose an experimental evaluation of a linear quadratic regulator (LQR) using a linearization of the full robot dynamics together with the contact constraints. The advantage of the controller is that it explicitly takes into account the coupling between the different joints to create optimal feedback controllers for whole-body control. We also propose a method to explicitly regulate other tasks of interest, such as the regulation of the center of mass of the robot or its angular momentum. In order to evaluate the performance of linear optimal control designs in a real-world scenario (model uncertainty, sensor noise, imperfect state estimation, etc), we test the controllers in a variety of tracking and balancing experiments on a torque controlled humanoid (e.g. balancing, split plane balancing, squatting, pushes while squatting, and balancing on a wheeled platform). The proposed control framework shows a reliable push recovery behavior competitive with more sophisticated balance controllers, rejecting impulses up to 11.7 Ns with peak forces of 650 N, with the added advantage of great computational simplicity. Furthermore, the controller is able to track squatting trajectories up to 1 Hz without relinearization, suggesting that the linearized dynamics is sufficient for significant ranges of motion.
AB - Humanoid robots operating in human environments require whole-body controllers that can offer precise tracking and well-defined disturbance rejection behavior. In this contribution, we propose an experimental evaluation of a linear quadratic regulator (LQR) using a linearization of the full robot dynamics together with the contact constraints. The advantage of the controller is that it explicitly takes into account the coupling between the different joints to create optimal feedback controllers for whole-body control. We also propose a method to explicitly regulate other tasks of interest, such as the regulation of the center of mass of the robot or its angular momentum. In order to evaluate the performance of linear optimal control designs in a real-world scenario (model uncertainty, sensor noise, imperfect state estimation, etc), we test the controllers in a variety of tracking and balancing experiments on a torque controlled humanoid (e.g. balancing, split plane balancing, squatting, pushes while squatting, and balancing on a wheeled platform). The proposed control framework shows a reliable push recovery behavior competitive with more sophisticated balance controllers, rejecting impulses up to 11.7 Ns with peak forces of 650 N, with the added advantage of great computational simplicity. Furthermore, the controller is able to track squatting trajectories up to 1 Hz without relinearization, suggesting that the linearized dynamics is sufficient for significant ranges of motion.
UR - http://www.scopus.com/inward/record.url?scp=84945179074&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84945179074&partnerID=8YFLogxK
U2 - 10.1109/HUMANOIDS.2014.7041387
DO - 10.1109/HUMANOIDS.2014.7041387
M3 - Conference contribution
AN - SCOPUS:84945179074
T3 - IEEE-RAS International Conference on Humanoid Robots
SP - 374
EP - 379
BT - 2014 IEEE-RAS International Conference on Humanoid Robots, Humanoids 2014
PB - IEEE Computer Society
Y2 - 18 November 2014 through 20 November 2014
ER -