On the Equivalence of Control Systems on the Orthogonal Group SO (4)
Ross M. Adams ∗ , Rory Biggs, Claudiu C. Remsing
Department of Mathematics (Pure and Applied) Rhodes University, Grahamstown 6140, South Africa
WSEAS International Conference on DYNAMICAL SYSTEMS and
CONTROL, Porto, Portugal, July 1–3, 2012
Outline
1 Introduction
2 The orthogonal group SO (4)
3 Equivalence
4 An illustrative example
Outline
1 Introduction
2 The orthogonal group SO (4)
3 Equivalence
4 An illustrative example
Problem statement
Left-invariant control affine systems on Lie groups.
Study the local geometry by introducing a natural equivalence relation.
Detached feedback equivalence and L-equivalence.
Classify, under L-equivalence, all homogeneous systems on
SO (4).
Left-invariant control affine systems
A left-invariant control affine system Σ = (G, Ξ) g ˙ = g Ξ(1, u) = g(A + u 1 B 1 + . . . u ` B ` ) where g ∈ G , u ∈ R ` and A, B 1 , . . . , B ` ∈ g.
The parameterization map Ξ(1, ·) : R ` → g is an injective affine map (i.e., B 1 , . . . , B ` are linearly independent).
The trace of Σ is Γ = im Ξ(1, ·) = A + hB 1 , . . . , B ` i.
Σ is homogeneous if A ∈ hB 1 , . . . , B ` i.
Equivalences
Let Σ = (G, Ξ) and Σ 0 = (G, Ξ 0 ).
Detached feedback equivalence
Σ and Σ 0 are (locally) detached feedback equivalent if there exist 1 ∈ N and 1 ∈ N 0 , and
a (local) diffeomorphism Φ = φ × ϕ : N × R ` → N 0 × R ` , φ(1) = 1, such that
T g φ · Ξ (g, u) = Ξ 0 (φ(g), ϕ(u)) for all g ∈ N and u ∈ R `
L-equivalence
Σ and Σ 0 are L-equivalent if there exists a Lie algebra automorphism ψ : g → g such that
ψ · Γ = Γ 0 .
Outline
1 Introduction
2 The orthogonal group SO (4)
3 Equivalence
4 An illustrative example
The orthogonal group SO (4)
The orthogonal group SO (4) = n
g ∈ GL (4, R ) : g > g = 1, det g = 1 o is a six-dimensional, non-commutative, semisimple, compact Lie group.
The Lie algebra
so (4) = n
A ∈ R 4×4 : A > + A = 0 o
.
Decomposition so (4) ∼ = so (3) ⊕ so (3)
Natural Basis
Isomorphism ς : so (3) ⊕ so (3) → so (4).
Induces a natural basis for so (4).
E 1 E 2 E 3 E 4 E 5 E 6
E 1 0 E 3 −E 2 0 0 0
E 2 −E 3 0 E 1 0 0 0
E 3 E 2 −E 1 0 0 0 0
E 4 0 0 0 0 E 6 −E 5
E 5 0 0 0 −E 6 0 E 4
E 6 0 0 0 E 5 −E 4 0
Automorphisms of so (4)
Lemma
Group of inner automorphisms Int (so (4)) =
ψ 1 0 0 ψ 2
: ψ 1 , ψ 2 ∈ SO (3)
.
Proposition
Aut (so (4)) is generated by Int (so (4)) and the automorphism ζ =
0 1 1 0
.
Outline
1 Introduction
2 The orthogonal group SO (4)
3 Equivalence
4 An illustrative example
Basic observations
Proposition
Let Γ, e Γ ⊂ so (4) and ψ ∈ Aut (so (4)). Then ψ · Γ = e Γ ⇐⇒ ψ · Γ ⊥ = e Γ ⊥ .
Any R ∈ SO (3) can be expressed as a product of rotations ρ 1 (θ), ρ 2 (θ) and ρ 3 (θ), denoted respectively
1 0 0
0 cos θ − sin θ 0 sin θ cos θ
,
cos θ 0 sin θ
0 1 0
− sin θ 0 cos θ
cos θ − sin θ 0 sin θ cos θ 0
0 0 1
.
Example
Let the trace of Σ be Γ =
D E 1 +
√
3E 3 + E 4 , 3E 2 + E 5 + E 6 E . Then
(ρ 2 ( π 3 ), 1) · Γ =
cos π 3 E 1 − sin π 3 E 3 +
√
3 (sin π 3 E 1 + cos π 3 E 3 ) + E 4 , 3E 2 + E 5 + E 6
= h2E 1 + E 4 , 3E 2 + E 5 + E 6 i . Also,
(1, ρ 1 (− π 4 )) · h2E 1 + E 4 , 3E 2 + E 5 + E 6 i
= D
E 1 + 1 2 E 4 , E 2 +
√ 2 3 E 5
E
= e Γ.
Therefore, Σ is L-equivalent to Σ e (with trace e Γ). Note that ψ = ρ 2 ( π 3 ), ρ 1 (− π 4 )
is such that ψ · Γ ⊥ = e Γ ⊥ .
One-input systems
Theorem
Any one-input system is L-equivalent to a system
Ξ 1 1 (1, u) = u 1 E 1
Ξ 1 2,α (1, u) = u 1 (E 1 + αE 4 )
for some 0 < α ≤ 1.
Proof
Let Σ be a one-input system.
Let A 1 = P 3
i=1 a i E i and A 2 = P 6
i=4 a i E i .
Then Γ 1 = hA 1 i, Γ 2 = hA 2 i, or Γ 3 = hA 1 + A 2 i.
Proof (cont.)
For Γ 1 , ∃ ψ = (ψ 1 , 1) ∈ Int(so (4)) such that ψ · Γ 1 = hE 1 i = Γ 1 1 . Similarly, there exists a ψ = (1, ψ 2 ) such that ψ · Γ 2 = hE 4 i.
Hence ζ · ψ · Γ 2 = hE 1 i = Γ 1 1 .
For Γ 3 , there exists a ψ = (ψ 1 , ψ 2 ) such that ψ · Γ 3 = hE 1 + αE 4 i for some α > 0.
If α ≤ 1, then ψ · Γ 3 = Γ 1 2,α .
If α > 1, then ζ · ψ · Γ 3 = hE 4 + αE 1 i =
E 1 + α 1 E 4
= Γ 1 2,
1 α.
Five-input systems
Corollary
Any five-input system is L-equivalent to a system
Ξ 5 1 (1, u) = u 1 E 2 + u 2 E 3 + u 3 E 4 + u 4 E 5 + u 6 E 6 Ξ 5 2,α (1, u) = u 1 E 2 + u 2 E 3 + u 3 E 5
+ u 4 E 6 + u 5 (αE 1 − E 4 )
for some 0 < α ≤ 1.
Any five-input system has full rank.
Two-input systems
Theorem
Any two-input system is L-equivalent to a system
Ξ 2 1 (1, u) = u 1 E 1 + u 2 E 2 Ξ 2 2 (1, u) = u 1 E 1 + u 2 E 4
Ξ 2 3,α (1, u) = u 1 E 1 + u 2 (E 2 + αE 5 ) Ξ 2 4,αβ (1, u) = u 1 (E 1 + αE 4 + βE 5 )
+ u 2 (E 2 + αE 5 )
for some α > 0, β ≥ 0.
Four-input systems
Corollary
Any four-input system is L-equivalent to a system
Ξ 4 1 (1, u) = u 1 E 3 + u 2 E 4 + u 3 E 5 + u 4 E 6 Ξ 4 2 (1, u) = u 1 E 2 + u 2 E 3 + u 5 E 5 + u 6 E 6 Ξ 4 3,α (1, u) = u 1 E 3 + u 2 E 4 + u 3 E 6
+ u 4 (αE 2 − E 5 )
Ξ 4 4,αβ (1, u) = u 1 E 3 + u 2 E 6 + u 3 βE 1 + αE 2
− E 5
+ u 4 (αE 1 − E 4 )
for some α > 0 and β ≥ 0.
Three-input systems
Theorem
Any three-input system is L-equivalent to a system
Ξ 3 1,αβ (1, u) = u 1 (E 1 + α 1 E 4 ) + u 2 (E 2 + α 2 E 5 ) + u 3 (E 3 + βE 6 )
Ξ 3 2,γ (1, u) = u 1 (E 1 + γE 4 ) + u 2 (E 2 + γ E 5 ) + u 3 (E 3 ± γE 6 )
Ξ 3 3,γ (1, u) = u 1 (E 1 + γE 4 ) + u 2 E 2 + u 3 E 6
for some α 1 ≥ α 2 ≥ |β| ≥ 0 and 0 ≤ γ ≤ 1. Here α 1 6= α 2 or
α 2 6= |β|.
Outline
1 Introduction
2 The orthogonal group SO (4)
3 Equivalence
4 An illustrative example
An illustrative example
Any system
Ξ γ (1, u) = γ 1 E 4 + u 1 (γ 2 E 1 + γ 3 E 2 + γ 4 E 3 )
+ u 2 (γ 5 E 3 + γ 6 E 4 ) + u 3 (γ 7 E 4 ) + u 4 (γ 8 E 5 )
is L-equivalent to the system
Ξ(1, e u) = u 1 E 2 + u 2 E 3 + u 3 E 4 + u 4 E 5 .
Here γ i > 0, i = 1, . . . , 8. The automorphism relating the traces of these systems is given by ψ = (ψ 1 , 1) , where
ψ 1 =
γ
3√
γ
22+γ
23− √ γ
2γ
22+γ
320
γ
2√
γ
22+γ
23γ
3√
γ
22+γ
320
0 0 1
.
An illustrative example
The corresponding feedback transformation ϕ, defined by ψ · Ξ γ (1, u) = Ξ (1, ϕ(u)) e
is given by
ϕ : u 7→
q
γ 2 2 + γ 3 2 0 0 0
γ 4 γ 5 0 0
0 γ 6 γ 7 0
0 0 0 γ 8
u +
0 0 γ 1
0
.
There exists a φ ∈ Aut(SO (4)) such that T 1 φ = ψ.
φ establishes a one-to-one correspondence between trajectories
of Ξ γ and Ξ. e
Concluding remark
Obtained a list of equivalence representatives for homogeneous systems on SO (4).
Attempt to obtain a classification of systems.
A classification can be obtained for the one-input case.
For the two-input and three-input cases the calculations become