Proportional Derivative (PD) Control on the Euclidean Group: Difference between revisions

From Murray Wiki
Jump to navigationJump to search
(htdb2wiki: creating page for 1995m_bm95b-cds.html)
 
(htdb2wiki: creating page for 1994n_bm95-ecc.html)
 
Line 2: Line 2:
| authors =  Francesco Bullo and Richard Murray  
| authors =  Francesco Bullo and Richard Murray  
| title = Proportional Derivative (PD) Control on the Euclidean Group
| title = Proportional Derivative (PD) Control on the Euclidean Group
| source = CDS Technical Report 95-010
| source = 1995 European Control Conference (Rome)
| year = 1995
| year = 1994
| type = CDS Technical Report
| type = Conference paper
| funding = NSF
| funding =  
| url = http://www.cds.caltech.edu/~murray/preprints/cds95-010.pdf
| url = http://www.cds.caltech.edu/~murray/preprints/bm95-ecc.pdf
| abstract =  
| abstract =  
In this paper we study the stabilization problem for control systems defined
In this paper we study the stabilization problem for control systems
on SE(3) (the special Euclidean group of rigid-body motions) and its
defined on SE(3), the Euclidean group of rigid--body motions.
subgroups.  Assuming one actuator is available for each degree of freedom,
Assuming one actuator is available for each degree of freedom, we
we exploit geometric properties of Lie groups (and corresponding Lie
exploit geometric properties of Lie groups (and corresponding Lie
algebras) to generalize the classical proportional derivative (PD) control
algebras) to generalize the classical PD control in a coordinate--free
in a coordinate-free way.  For the SO(3) case, the compactness of the group
way.  For the SO(3) case, the compactness of the group gives rise to a
gives rise to a natural metric structure and to a natural choice of
natural metric structure and to a natural choice of preferred control
preferred control direction: an optimal (in the sense of geodesic) solution
direction: an optimal (in the sense of geodesic) solution is given to
is given to the attitude control problem.  In the SE(3) case, no natural
the attitude control problem.  In the SE(3) case, no natural metric is
metric is uniquely defined, so that more freedom is left in the control
uniquely defined, so that more freedom is left in the control design.
design. Different formulations of PD feedback can be adopted by extending
Different formulations of PD feedback can be adopted by extending the
the SO(3) approach to the whole of SE(3) or by breaking the problem into a
SO(3) approach to the whole of SE(3) or by breaking the problem into a
control problem on SO(3) x R^3.  For the simple SE(2) case, simulations are
control problem on SO(3) x R^3.  For the simple SE(2) case,
reported to illustrate the behavior of the different choices.  We also
simulations are reported to illustrate the behavior of the different
discuss the trajectory tracking problem and show how to reduce it to a
choices.  Finally, we discuss the trajectory tracking problem and show
stabilization problem, mimicking the usual approach in R^n.  Finally,
how to reduce it to a stabilization problem, mimicking the usual
regarding the case of underactuated control systems, we derive linear and
approach in R^n.
homogeneous approximating vector fields for standard systems on SO(3) and
SE(3).
| flags = NoRequest
| flags = NoRequest
| tag = bm95b-cds
| tag = bm95-ecc
| id = 1995m
| id = 1994n
}}
}}

Latest revision as of 06:20, 15 May 2016


Francesco Bullo and Richard Murray
1995 European Control Conference (Rome)

In this paper we study the stabilization problem for control systems defined on SE(3), the Euclidean group of rigid--body motions. Assuming one actuator is available for each degree of freedom, we exploit geometric properties of Lie groups (and corresponding Lie algebras) to generalize the classical PD control in a coordinate--free way. For the SO(3) case, the compactness of the group gives rise to a natural metric structure and to a natural choice of preferred control direction: an optimal (in the sense of geodesic) solution is given to the attitude control problem. In the SE(3) case, no natural metric is uniquely defined, so that more freedom is left in the control design. Different formulations of PD feedback can be adopted by extending the SO(3) approach to the whole of SE(3) or by breaking the problem into a control problem on SO(3) x R^3. For the simple SE(2) case, simulations are reported to illustrate the behavior of the different choices. Finally, we discuss the trajectory tracking problem and show how to reduce it to a stabilization problem, mimicking the usual approach in R^n.