| タイトル | An innovations approach to decoupling of multibody dynamics and control |
| 本文（外部サイト） | http://hdl.handle.net/2060/19900013678 |
| 著者（英） | Rodriguez, G. |
| 著者所属（英） | Jet Propulsion Lab., California Inst. of Tech. |
| 発行日 | 1989-12-15 |
| 言語 | eng |
| 内容記述 | The problem of hinged multibody dynamics is solved using an extension of the innovations approach of linear filtering and prediction theory to the problem of mechanical system modeling and control. This approach has been used quite effectively to diagonalize the equations for filtering and prediction for linear state space systems. It has similar advantages in the study of dynamics and control of multibody systems. The innovations approach advanced here consists of expressing the equations of motion in terms of two closely related processes: (1) the innovations process e, a sequence of moments, obtained from the applied moments T by means of a spatially recursive Kalman filter that goes from the tip of the manipulator to its base; (2) a residual process, a sequence of velocities, obtained from the joint-angle velocities by means of an outward smoothing operations. The innovations e and the applied moments T are related by means of the relationships e = (I - L)T and T = (I + K)e. The operation (I - L) is a causal lower triangular matrix which is generated by a spatially recursive Kalman filter and the corresponding discrete-step Riccati equation. Hence, the innovations and the applied moments can be obtained from each other by means of a causal operation which is itself casually invertible. |
| NASA分類 | COMPUTER PROGRAMMING AND SOFTWARE |
| レポートNO | 90N22994 |
| 権利 | No Copyright |
| URI | https://repository.exst.jaxa.jp/dspace/handle/a-is/137782 |
|
