Lee, Jongwoo Kim, Jung Hoon Oh, Yonghwan 2024-01-19T11:38:13Z 2024-01-19T11:38:13Z 2022-03-07 2016 https://pubs.kist.re.kr/handle/201004/114985 Despite successful demonstrations of outdoor walking by a few biped robots, performance measure for such robots has not been formally defined yet. The performance measure should be adequately defined by which one can evaluate how well the robot keeps from falling in the presence of disturbance. If such performance measure is suitably determined, designing a sort of optimal controller for stable outside walking would be possible. This paper firstly suggests to adopt the l(infinity)-induced norm defined on the linearized Poincare map as a novel performance measure for biped robots, in the presence of bounded persistent disturbance. In order to validate the measure, a nonlinear dynamic simulation is conducted by extending an existing simple model to walk on rough terrain, of which height variance is bounded by some maximum value. The measure exploited for the model is compared with the numerical result obtained from nonlinear dynamic simulation. Finally, an example of application of this measure is provided, which successfully predicts whether the system could overcome upcoming terrain roughness. English IEEE A Novel Performance Measure for Biped Robots against Bounded Persistent Disturbances Conference 1 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp.5805 - 5812 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 5805 5812 US Daejeon, SOUTH KOREA 2016-10-09 2016 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS 2016) 000391921705125 2-s2.0-85006380596 Proceedings Paper