Abstract | ||
---|---|---|
We propose a general but simple bipedal walking control architecture that incorporates intuitive control and learning algorithms. The learning algorithm is mainly used to generate the key parameters for the swing leg. The intuitive control is used to maintain the height and body posture. Based on the proposed architecture, a control algorithm is constructed and applied to a planar biped and a 3D biped. By applying an appropriate local speed control mechanism, we demonstrate that the bipeds can successfully achieve walking of 100 seconds within a reasonable number of trials. No dynamic models or nominal joint trajectory data are required for the implementations. |
Year | DOI | Venue |
---|---|---|
2000 | 10.1109/ROBOT.2000.845353 | ICRA |
Keywords | Field | DocType |
biomechanics,learning artificial intelligence,torso,taxonomy,algorithm design and analysis,leg,speed control | Control algorithm,Architecture,Control theory,Simulation,Control engineering,Body posture,Dynamic models,Engineering,Trajectory,Swing,Electronic speed control | Conference |
Volume | Issue | ISSN |
4 | 1 | 1050-4729 |
Citations | PageRank | References |
14 | 1.64 | 9 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Chee-Meng Chew | 1 | 375 | 40.58 |
gill a pratt | 2 | 721 | 116.12 |