已知流形： E = ( a ⊞ b ) ⊟ ( c ⊞ d ) ， a , c ∈ S O ( 3 ) , b , d ∈ R 3 E=(a \boxplus b) \boxminus ( c \boxplus d)，a,c \in \mathit{SO(3)}, b,d \in \mathbb{R}^3 E=(a⊞b)⊟(c⊞d)，a,c∈SO(3),b,d∈R3，求E对b和E对d的偏导数
（1）E对b的偏导数的推导：
E
=
(
a
⊞
b
)
⊟
(
c
⊞
d
)
→
E
=
L
o
g
(
(
c
E
x
p
(
d
)
)
−
1
a
E
x
p
(
b
)
)
→
E
x
p
(
E
)
=
(
c
E
x
p
(
d
)
)
−
1
a
E
x
p
(
b
)
→
E
x
p
(
E
)
=
E
x
p
(
−
d
)
c
−
1
a
E
x
p
(
b
)
→
E
x
p
(
E
+
Δ
E
)
=
E
x
p
(
−
d
)
c
−
1
a
E
x
p
(
b
+
Δ
b
)
→
E
x
p
(
E
)
E
x
p
[
A
(
E
)
T
Δ
E
]
=
E
x
p
(
−
d
)
c
−
1
a
E
x
p
(
b
)
E
x
p
(
A
(
b
)
T
Δ
b
)
→
E
x
p
(
E
)
E
x
p
[
A
(
E
)
T
Δ
E
]
=
E
x
p
(
E
)
E
x
p
(
A
(
b
)
T
Δ
b
)
→
E
x
p
[
A
(
E
)
T
Δ
E
]
=
E
x
p
(
A
(
b
)
T
Δ
b
)
→
A
(
E
)
T
Δ
E
=
A
(
b
)
T
Δ
b
→
Δ
E
Δ
b
=
A
(
E
)
−
T
A
(
b
)
T
→
∂
E
∂
b
=
A
(
E
)
−
T
A
(
−
b
)
\begin{align} E & =(a \boxplus b) \boxminus ( c \boxplus d) \\ \to E & = Log( (c Exp(d)) ^ {-1} a Exp(b)) \\ \to Exp(E) & = (c Exp(d)) ^ {-1} a Exp(b) \\ \to Exp(E) & = Exp(-d) c ^ {-1} a Exp(b) \\ \to Exp(E+\Delta E) & = Exp(-d) c ^ {-1} a Exp(b+ \Delta b) \\ \to Exp(E)Exp[A(E) ^ T \Delta E] & = Exp(-d) c ^ {-1} a Exp(b) Exp(A(b) ^ T \Delta b) \\ \to Exp(E)Exp[A(E) ^ T \Delta E] & = Exp(E) Exp(A(b) ^ T \Delta b) \\ \to Exp[A(E) ^ T \Delta E] & = Exp(A(b) ^ T \Delta b) \\ \to A(E) ^ T \Delta E & = A(b) ^ T \Delta b \\ \to \frac {\Delta E} {\Delta b} & = A(E) ^ {-T} A(b) ^ T \\ \to \frac {\partial E} {\partial b} & = A(E) ^ {-T} A(-b) \end{align}
E→E→Exp(E)→Exp(E)→Exp(E+ΔE)→Exp(E)Exp[A(E)TΔE]→Exp(E)Exp[A(E)TΔE]→Exp[A(E)TΔE]→A(E)TΔE→ΔbΔE→∂b∂E=(a⊞b)⊟(c⊞d)=Log((cExp(d))−1aExp(b))=(cExp(d))−1aExp(b)=Exp(−d)c−1aExp(b)=Exp(−d)c−1aExp(b+Δb)=Exp(−d)c−1aExp(b)Exp(A(b)TΔb)=Exp(E)Exp(A(b)TΔb)=Exp(A(b)TΔb)=A(b)TΔb=A(E)−TA(b)T=A(E)−TA(−b)
（2）E对d的偏导数的推导：
E
=
(
a
⊞
b
)
⊟
(
c
⊞
d
)
→
E
=
L
o
g
(
(
c
E
x
p
(
d
)
)
−
1
a
E
x
p
(
b
)
)
→
E
x
p
(
E
)
=
(
c
E
x
p
(
d
)
)
−
1
a
E
x
p
(
b
)
→
E
x
p
(
E
)
=
E
x
p
(
−
d
)
c
−
1
a
E
x
p
(
b
)
→
E
x
p
(
E
+
Δ
E
)
=
E
x
p
(
−
d
−
Δ
d
)
c
−
1
a
E
x
p
(
b
)
→
E
x
p
(
E
)
E
x
p
[
A
(
E
)
T
Δ
E
]
=
E
x
p
(
−
d
)
E
x
p
[
A
(
−
d
)
T
(
−
Δ
d
)
]
c
−
1
a
E
x
p
(
b
)
→
E
x
p
[
A
(
E
)
T
Δ
E
]
=
(
c
−
1
a
E
x
p
(
b
)
)
−
1
E
x
p
[
A
(
−
d
)
T
(
−
Δ
d
)
]
c
−
1
a
E
x
p
(
b
)
→
E
x
p
[
A
(
E
)
T
Δ
E
]
=
E
x
p
[
(
c
−
1
a
E
x
p
(
b
)
)
−
1
A
(
−
d
)
T
(
−
Δ
d
)
]
→
A
(
E
)
T
Δ
E
=
(
c
−
1
a
E
x
p
(
b
)
)
−
1
A
(
−
d
)
T
(
−
Δ
d
)
→
A
(
E
)
T
Δ
E
=
−
(
c
−
1
a
E
x
p
(
b
)
)
−
1
A
(
d
)
Δ
d
→
A
(
E
)
T
Δ
E
=
−
E
x
p
(
−
b
)
a
−
1
c
A
(
d
)
Δ
d
→
Δ
E
Δ
d
=
−
A
(
E
)
−
T
E
x
p
(
−
b
)
a
−
1
c
A
(
d
)
→
∂
E
∂
d
=
−
A
(
E
)
−
T
E
x
p
(
−
b
)
a
−
1
c
A
(
d
)
\begin{align} E & =(a \boxplus b) \boxminus ( c \boxplus d) \\ \to E & = Log( (c Exp(d)) ^ {-1} a Exp(b)) \\ \to Exp(E) & = (c Exp(d)) ^ {-1} a Exp(b) \\ \to Exp(E) & = Exp(-d) c ^ {-1} a Exp(b) \\ \to Exp(E + \Delta E) & = Exp(-d- \Delta d) c ^ {-1} a Exp(b) \\ \to Exp(E) Exp[ A(E) ^ T \Delta E] & = Exp(-d)Exp[A(-d) ^ T (- \Delta d) ] c ^ {-1} a Exp(b) \\ \to Exp[ A(E) ^ T \Delta E] & = {(c ^ {-1} a Exp(b) )} ^ {-1} Exp[A(-d) ^ T (- \Delta d) ] c ^ {-1} a Exp(b) \\ \to Exp[ A(E) ^ T \Delta E] & = Exp[(c ^ {-1} a Exp(b) ) ^ {-1}A(-d) ^ T (- \Delta d) ] \\ \to A(E) ^ T \Delta E & = (c ^ {-1} a Exp(b) ) ^ {-1}A(-d) ^ T (- \Delta d) \\ \to A(E) ^ T \Delta E & = - (c ^ {-1} a Exp(b) ) ^ {-1} A(d) \Delta d \\ \to A(E) ^ T \Delta E & = - Exp(-b) a ^ {-1} c A(d) \Delta d \\ \to \frac {\Delta E} { \Delta d }& = - A(E) ^ {-T} Exp(-b) a ^ {-1} c A(d) \\ \to \frac {\partial E} { \partial d }& = - A(E) ^ {-T} Exp(-b) a ^ {-1} c A(d) \\ \end{align}
E→E→Exp(E)→Exp(E)→Exp(E+ΔE)→Exp(E)Exp[A(E)TΔE]→Exp[A(E)TΔE]→Exp[A(E)TΔE]→A(E)TΔE→A(E)TΔE→A(E)TΔE→ΔdΔE→∂d∂E=(a⊞b)⊟(c⊞d)=Log((cExp(d))−1aExp(b))=(cExp(d))−1aExp(b)=Exp(−d)c−1aExp(b)=Exp(−d−Δd)c−1aExp(b)=Exp(−d)Exp[A(−d)T(−Δd)]c−1aExp(b)=(c−1aExp(b))−1Exp[A(−d)T(−Δd)]c−1aExp(b)=Exp[(c−1aExp(b))−1A(−d)T(−Δd)]=(c−1aExp(b))−1A(−d)T(−Δd)=−(c−1aExp(b))−1A(d)Δd=−Exp(−b)a−1cA(d)Δd=−A(E)−TExp(−b)a−1cA(d)=−A(E)−TExp(−b)a−1cA(d)
997
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
