本文仅供学习使用,总结很多本现有讲述运动学或动力学书籍后的总结,从矢量的角度进行分析,方法比较传统,但更易理解,并且现有的看似抽象方法,两者本质上并无不同。
2024年底本人学位论文发表后方可摘抄
若有帮助请引用
本文参考:
.
食用方法
如何表达刚体在空间中的位置与姿态
姿态参数如何表达?不同表达方式直接的转换关系?
旋转矩阵?转换矩阵?有什么意义和性质?转置代表什么?
如何表示连续变换?——与RPY有关
齐次坐标的意义——简化公式?
务必自己推导全部公式,并理解每个符号的含义
机构运动学与动力学分析与建模 Ch00-2质量刚体的在坐标系下运动Part2
- 3.4 欧拉四元数变换
- 3.4.1 四元数的数学性质
- 3.4.2 四元数与轴角的转换
- 3.4.3 四元数旋转矢量
- 3.5 欧拉角(ZYX变换)与RPY角 Euler Angles
- 3.5.1 欧拉角与四元数的转换
- 3.5.2 欧拉角与轴角的转换
- 3.6 连续转动
- 3.7 齐次矩阵的表达
- 3.8 点与向量在不同坐标系下的表达
3.4 欧拉四元数变换
同样基于罗德里格旋转公式,定义四个欧拉参数为:
q
⃗
=
[
s
v
⃗
]
=
[
cos
θ
2
→
s
c
a
l
e
p
a
r
t
v
⃗
sin
θ
2
→
v
e
c
t
o
r
p
a
r
t
]
=
[
cos
θ
2
v
1
sin
θ
2
v
2
sin
θ
2
v
3
sin
θ
2
]
=
[
q
1
q
2
q
3
q
4
]
\vec{q}=\left[ \begin{array}{c} s\\ \vec{v}\\ \end{array} \right] =\left[ \begin{matrix} \cos \frac{\theta}{2}& \rightarrow scale\,\,part\\ \vec{v}\sin \frac{\theta}{2}& \rightarrow vector\,\,part\\ \end{matrix} \right] =\left[ \begin{array}{c} \cos \frac{\theta}{2}\\ v_1\sin \frac{\theta}{2}\\ v_2\sin \frac{\theta}{2}\\ v_3\sin \frac{\theta}{2}\\ \end{array} \right] =\left[ \begin{array}{c} q_1\\ q_2\\ q_3\\ q_4\\ \end{array} \right]
q=[sv]=[cos2θvsin2θ→scalepart→vectorpart]=
cos2θv1sin2θv2sin2θv3sin2θ
=
q1q2q3q4
3.4.1 四元数的数学性质
- 归一性 : q ⃗ T q ⃗ = ∑ i = 1 n q i 2 = 1 \vec{q}^{\mathrm{T}}\vec{q}=\sum_{i=1}^n{{q_{\mathrm{i}}}^2}=1 qTq=∑i=1nqi2=1
- 四元数的正交性(逆) : q ⃗ T q ⃗ = 1 ⇒ q ⃗ T = q ⃗ − 1 \vec{q}^{\mathrm{T}}\vec{q}=1\Rightarrow \vec{q}^{\mathrm{T}}=\vec{q}^{-1} qTq=1⇒qT=q−1
- 四元数的转置(共轭)——旋转轴不变,旋转角相反 : q ⃗ T = q ⃗ − 1 = [ cos ( − θ 2 ) v 1 sin ( − θ 2 ) v 2 sin ( − θ 2 ) v 3 sin ( − θ 2 ) ] = [ cos θ 2 − v 1 sin θ 2 − v 2 sin θ 2 − v 3 sin θ 2 ] = [ s − v ⃗ ] \vec{q}^{\mathrm{T}}=\vec{q}^{-1}=\left[ \begin{array}{c} \cos \left( \frac{-\theta}{2} \right)\\ v_1\sin \left( \frac{-\theta}{2} \right)\\ v_2\sin \left( \frac{-\theta}{2} \right)\\ v_3\sin \left( \frac{-\theta}{2} \right)\\ \end{array} \right] =\left[ \begin{array}{c} \cos \frac{\theta}{2}\\ -v_1\sin \frac{\theta}{2}\\ -v_2\sin \frac{\theta}{2}\\ -v_3\sin \frac{\theta}{2}\\ \end{array} \right] =\left[ \begin{array}{c} s\\ -\vec{v}\\ \end{array} \right] qT=q−1= cos(2−θ)v1sin(2−θ)v2sin(2−θ)v3sin(2−θ) = cos2θ−v1sin2θ−v2sin2θ−v3sin2θ =[s−v]
- 四元数的乘法 :
q
⃗
1
⋅
q
⃗
2
=
[
s
1
v
⃗
1
]
⋅
[
s
2
v
⃗
2
]
=
[
s
1
s
2
−
v
⃗
1
T
v
⃗
2
s
1
v
⃗
2
+
s
2
v
⃗
1
+
v
⃗
1
×
v
⃗
2
]
=
[
s
1
−
v
⃗
1
T
v
⃗
1
s
1
E
+
v
⃗
~
1
]
⏟
L
(
q
1
)
[
s
2
v
⃗
2
]
=
[
s
2
−
v
⃗
2
T
v
⃗
2
s
2
E
−
v
⃗
~
2
]
⏟
R
(
q
2
)
[
s
1
v
⃗
1
]
\begin{split} \vec{q}_1\cdot \vec{q}_2&=\left[ \begin{array}{c} s_1\\ \vec{v}_1\\ \end{array} \right] \cdot \left[ \begin{array}{c} s_2\\ \vec{v}_2\\ \end{array} \right] =\left[ \begin{array}{c} s_1s_2-{\vec{v}_1}^{\mathrm{T}}\vec{v}_2\\ s_1\vec{v}_2+s_2\vec{v}_1+\vec{v}_1\times \vec{v}_2\\ \end{array} \right] \\ &=\begin{array}{c} \underbrace{\left[ \begin{matrix} s_1& -{\vec{v}_1}^{\mathrm{T}}\\ \vec{v}_1& s_1E+\tilde{\vec{v}}_1\\ \end{matrix} \right] }\\ L\left( q_1 \right)\\ \end{array}\left[ \begin{array}{c} s_2\\ \vec{v}_2\\ \end{array} \right] =\begin{array}{c} \underbrace{\left[ \begin{matrix} s_2& -{\vec{v}_2}^{\mathrm{T}}\\ \vec{v}_2& s_2E-\tilde{\vec{v}}_2\\ \end{matrix} \right] }\\ R\left( q_2 \right)\\ \end{array}\left[ \begin{array}{c} s_1\\ \vec{v}_1\\ \end{array} \right] \end{split}
q1⋅q2=[s1v1]⋅[s2v2]=[s1s2−v1Tv2s1v2+s2v1+v1×v2]=
[s1v1−v1Ts1E+v~1]L(q1)[s2v2]=
[s2v2−v2Ts2E−v~2]R(q2)[s1v1]
其中: L ( q 1 T ) = L ( q 1 ) T L\left( {q_1}^{\mathrm{T}} \right) =L\left( q_1 \right) ^{\mathrm{T}} L(q1T)=L(q1)T, R ( q 1 T ) = R ( q 1 ) T R\left( {q_1}^{\mathrm{T}} \right) =R\left( q_1 \right) ^{\mathrm{T}} R(q1T)=R(q1)T。 - 四元数的同一性 :
当 θ = 0 \theta =0 θ=0时: q ⃗ ∣ θ = 0 = [ 1 0 ⃗ ] \left. \vec{q} \right|_{\theta =0}=\left[ \begin{array}{c} 1\\ \vec{0}\\ \end{array} \right] q∣θ=0=[10]
根据上述定义,可将轴角变换的旋转矩阵
[
Q
B
A
]
\left[ Q_{\mathrm{B}}^{A} \right]
[QBA] 改写为:
[
Q
B
A
]
=
[
1
−
2
q
3
2
−
2
q
4
2
2
(
q
2
q
3
−
q
1
q
4
)
2
(
q
2
q
4
+
q
1
q
3
)
2
(
q
2
q
3
+
q
1
q
4
)
1
−
2
q
2
2
−
2
q
4
2
2
(
q
3
q
4
−
q
1
q
2
)
2
(
q
2
q
4
−
q
1
q
3
)
2
(
q
3
q
4
+
q
1
q
2
)
1
−
2
q
2
2
−
2
q
3
2
]
=
[
2
q
1
2
+
2
q
2
2
−
1
2
(
q
2
q
3
−
q
1
q
4
)
2
(
q
2
q
4
+
q
1
q
3
)
2
(
q
2
q
3
+
q
1
q
4
)
2
q
1
2
+
2
q
3
2
−
1
2
(
q
3
q
4
−
q
1
q
2
)
2
(
q
2
q
4
−
q
1
q
3
)
2
(
q
3
q
4
+
q
1
q
2
)
2
q
1
2
+
2
q
4
2
−
1
]
\left[ Q_{\mathrm{B}}^{A} \right] =\left[ \begin{matrix} 1-2{q_3}^2-2{q_4}^2& 2\left( q_2q_3-q_1q_4 \right)& 2\left( q_2q_4+q_1q_3 \right)\\ 2\left( q_2q_3+q_1q_4 \right)& 1-2{q_2}^2-2{q_4}^2& 2\left( q_3q_4-q_1q_2 \right)\\ 2\left( q_2q_4-q_1q_3 \right)& 2\left( q_3q_4+q_1q_2 \right)& 1-2{q_2}^2-2{q_3}^2\\ \end{matrix} \right] =\left[ \begin{matrix} 2{q_1}^2+2{q_2}^2-1& 2\left( q_2q_3-q_1q_4 \right)& 2\left( q_2q_4+q_1q_3 \right)\\ 2\left( q_2q_3+q_1q_4 \right)& 2{q_1}^2+2{q_3}^2-1& 2\left( q_3q_4-q_1q_2 \right)\\ 2\left( q_2q_4-q_1q_3 \right)& 2\left( q_3q_4+q_1q_2 \right)& 2{q_1}^2+2{q_4}^2-1\\ \end{matrix} \right]
[QBA]=
1−2q32−2q422(q2q3+q1q4)2(q2q4−q1q3)2(q2q3−q1q4)1−2q22−2q422(q3q4+q1q2)2(q2q4+q1q3)2(q3q4−q1q2)1−2q22−2q32
=
2q12+2q22−12(q2q3+q1q4)2(q2q4−q1q3)2(q2q3−q1q4)2q12+2q32−12(q3q4+q1q2)2(q2q4+q1q3)2(q3q4−q1q2)2q12+2q42−1
进而定义矩阵:
B
3
×
4
=
[
−
q
2
q
1
−
q
4
q
3
−
q
3
q
4
q
1
−
q
2
−
q
4
−
q
3
q
2
q
1
]
B
ˉ
3
×
4
=
[
−
q
2
q
1
q
4
−
q
3
−
q
3
−
q
4
q
1
q
2
−
q
4
q
3
−
q
2
q
1
]
\begin{split} B_{3\times 4}&=\left[ \begin{array}{cccc} -q_2& q_1& -q_4& q_3\\ -q_3& q_4& q_1& -q_2\\ -q_4& -q_3& q_2& q_1\\ \end{array} \right] \\ \bar{B}_{3\times 4}&=\left[ \begin{array}{cccc} -q_2& q_1& q_4& -q_3\\ -q_3& -q_4& q_1& q_2\\ -q_4& q_3& -q_2& q_1\\ \end{array} \right] \end{split}
B3×4Bˉ3×4=
−q2−q3−q4q1q4−q3−q4q1q2q3−q2q1
=
−q2−q3−q4q1−q4q3q4q1−q2−q3q2q1
则有:
[
Q
B
A
]
=
B
3
×
4
B
ˉ
3
×
4
T
\left[ Q_{\mathrm{B}}^{A} \right] =B_{3\times 4}{\bar{B}_{3\times 4}}^{\mathrm{T}}
[QBA]=B3×4Bˉ3×4T
上述矩阵具有如下性质:
B
3
×
4
B
3
×
4
T
=
B
ˉ
3
×
4
B
ˉ
3
×
4
T
=
E
B
3
×
4
T
B
3
×
4
=
B
ˉ
3
×
4
T
B
ˉ
3
×
4
=
E
4
×
4
−
q
⃗
⋅
q
⃗
T
B
ˉ
3
×
4
⋅
q
⃗
=
0
⃗
B_{3\times 4}{B_{3\times 4}}^{\mathrm{T}}=\bar{B}_{3\times 4}{\bar{B}_{3\times 4}}^{\mathrm{T}}=E \\ {B_{3\times 4}}^{\mathrm{T}}B_{3\times 4}={\bar{B}_{3\times 4}}^{\mathrm{T}}\bar{B}_{3\times 4}=E_{4\times 4}-\vec{q}\cdot \vec{q}^{\mathrm{T}} \\ \bar{B}_{3\times 4}\cdot \vec{q}=\vec{0}
B3×4B3×4T=Bˉ3×4Bˉ3×4T=EB3×4TB3×4=Bˉ3×4TBˉ3×4=E4×4−q⋅qTBˉ3×4⋅q=0
因此,若已知旋转矩阵:
[
Q
B
A
]
=
[
q
11
q
12
q
13
q
21
q
22
q
23
q
31
q
23
q
33
]
\left[ Q_{\mathrm{B}}^{A} \right] =\left[ \begin{matrix} q_{11}& q_{12}& q_{13}\\ q_{21}& q_{22}& q_{23}\\ q_{31}& q_{23}& q_{33}\\ \end{matrix} \right]
[QBA]=
q11q21q31q12q22q23q13q23q33
,则可求解四元数参数为:
[ q 1 q 2 q 3 q 4 ] = [ 1 2 q 11 + q 22 + q 33 + 1 q 32 − q 23 4 q 1 q 13 − q 31 4 q 1 q 21 − q 12 4 q 1 ] \left[ \begin{array}{c} q_1\\ q_2\\ q_3\\ q_4\\ \end{array} \right] =\left[ \begin{array}{c} \frac{1}{2}\sqrt{q_{11}+q_{22}+q_{33}+1}\\ \frac{q_{32}-q_{23}}{4q_1}\\ \frac{q_{13}-q_{31}}{4q_1}\\ \frac{q_{21}-q_{12}}{4q_1}\\ \end{array} \right] q1q2q3q4 = 21q11+q22+q33+14q1q32−q234q1q13−q314q1q21−q12
3.4.2 四元数与轴角的转换
- 四元数转换为轴角
[ θ v 1 v 2 v 3 ] = [ 2 a r c cos ( q 1 ) q 2 sin θ 2 q 3 sin θ 2 q 4 sin θ 2 ] \left[ \begin{array}{c} \theta\\ v_1\\ v_2\\ v_3\\ \end{array} \right] =\left[ \begin{array}{c} 2\mathrm{arc}\cos \left( q_1 \right)\\ \frac{q_2}{\sin \frac{\theta}{2}}\\ \frac{q_3}{\sin \frac{\theta}{2}}\\ \frac{q_4}{\sin \frac{\theta}{2}}\\ \end{array} \right] θv1v2v3 = 2arccos(q1)sin2θq2sin2θq3sin2θq4 - 轴角转换为四元数
[ q 1 q 2 q 3 q 4 ] = [ cos θ 2 v 1 sin θ 2 v 2 sin θ 2 v 3 sin θ 2 ] \left[ \begin{array}{c} q_1\\ q_2\\ q_3\\ q_4\\ \end{array} \right] =\left[ \begin{array}{c} \cos \frac{\theta}{2}\\ v_1\sin \frac{\theta}{2}\\ v_2\sin \frac{\theta}{2}\\ v_3\sin \frac{\theta}{2}\\ \end{array} \right] q1q2q3q4 = cos2θv1sin2θv2sin2θv3sin2θ
3.4.3 四元数旋转矢量
对于任意矢量
R
⃗
F
\vec{R}^F
RF,可通过上述四元数矢量进行旋转变化:
[
0
R
⃗
′
F
]
=
q
⃗
F
⋅
[
0
R
⃗
F
]
⋅
(
q
⃗
F
)
−
1
=
q
⃗
F
⋅
[
0
R
⃗
F
]
⋅
(
q
⃗
F
)
T
=
L
(
q
)
R
(
q
)
T
[
0
R
⃗
F
]
=
R
(
q
)
T
L
(
q
)
[
0
R
⃗
F
]
\left[ \begin{array}{c} 0\\ {\vec{R}^{\prime}}^F\\ \end{array} \right] =\vec{q}^F\cdot \left[ \begin{array}{c} 0\\ \vec{R}^F\\ \end{array} \right] \cdot \left( \vec{q}^F \right) ^{-1}=\vec{q}^F\cdot \left[ \begin{array}{c} 0\\ \vec{R}^F\\ \end{array} \right] \cdot \left( \vec{q}^F \right) ^{\mathrm{T}} \\ =L\left( q \right) R\left( q \right) ^{\mathrm{T}}\left[ \begin{array}{c} 0\\ \vec{R}^F\\ \end{array} \right] =R\left( q \right) ^{\mathrm{T}}L\left( q \right) \left[ \begin{array}{c} 0\\ \vec{R}^F\\ \end{array} \right]
[0R′F]=qF⋅[0RF]⋅(qF)−1=qF⋅[0RF]⋅(qF)T=L(q)R(q)T[0RF]=R(q)TL(q)[0RF]
同理也可以将上述视为矢量的坐标系变换,其转换矩阵拓展为
4
×
4
4\times 4
4×4,进而有转换矩阵
[
Q
B
A
]
4
×
4
T
\left[ Q_{\mathrm{B}}^{A} \right]^{\mathrm{T}} _{4\times 4}
[QBA]4×4T:
[
Q
B
A
]
4
×
4
=
L
(
q
)
R
(
q
)
T
=
R
(
q
)
T
L
(
q
)
\left[ Q_{\mathrm{B}}^{A} \right] _{4\times 4}=L\left( q \right) R\left( q \right) ^{\mathrm{T}}=R\left( q \right) ^{\mathrm{T}}L\left( q \right)
[QBA]4×4=L(q)R(q)T=R(q)TL(q)
此时,则有:
[
0
R
⃗
B
]
=
[
Q
B
A
]
4
×
4
T
[
0
R
⃗
A
]
\left[ \begin{array}{c} 0\\ \vec{R}^B\\ \end{array} \right] =\left[ Q_{\mathrm{B}}^{A} \right]^{\mathrm{T}} _{4\times 4}\left[ \begin{array}{c} 0\\ \vec{R}^A\\ \end{array} \right]
[0RB]=[QBA]4×4T[0RA]
3.5 欧拉角(ZYX变换)与RPY角 Euler Angles
欧拉角是一种较为原始的旋转表示方式,在实际的算法运用过程中,除了描述已知姿态的刚体角度外,在实际计算中,效果很差(具有奇异性、高度非线性、计算复杂)。因此,并不推荐用欧拉角来描述转换矩阵。本节仅对部分重点内容进行介绍。
欧拉角(ZYX变换)的旋转变换描述为:绕固定坐标系的基矢量 k ⃗ F \vec{k}^{F} kF回转 γ \gamma γ,得到新标架 { F 1 : ( i ⃗ 1 F , j ⃗ 1 F , k ⃗ 1 F ) } \left\{ F_1:\left( \vec{i}_{1}^{F},\vec{j}_{1}^{F},\vec{k}_{1}^{F} \right) \right\} {F1:(i1F,j1F,k1F)};再绕基矢量 j ⃗ 1 F \vec{j}_{1}^{F} j1F回转 β \beta β,得到新标架 { F 2 : ( i ⃗ 2 F , j ⃗ 2 F , k ⃗ 2 F ) } \left\{ F_2:\left( \vec{i}_{2}^{F},\vec{j}_{2}^{F},\vec{k}_{2}^{F} \right) \right\} {F2:(i2F,j2F,k2F)};最后绕基矢量 i ⃗ 2 F \vec{i}_{2}^{F} i2F回转 α \alpha α,得到新标架 { F 3 : ( i ⃗ 3 F , j ⃗ 3 F , k ⃗ 3 F ) } \left\{ F_3:\left( \vec{i}_{3}^{F},\vec{j}_{3}^{F},\vec{k}_{3}^{F} \right) \right\} {F3:(i3F,j3F,k3F)}为最终的变换姿态(运动坐标系下连续转动,右乘)。因此对于多次连续转动而言有:
[ Q M F ] = [ Q F 1 F ( k ⃗ F , γ ) ] [ Q F 2 F 1 ( j ⃗ 1 F , β ) ] [ Q F 3 ( M ) F 2 ( i ⃗ 2 F , α ) ] ] \left[ Q_{\mathrm{M}}^{F} \right] =\left[ Q_{\mathrm{F}_1}^{F}\left( \vec{k}^F,\gamma \right) \right] \left[ Q_{\mathrm{F}_2}^{F_1}\left( \vec{j}_{1}^{F},\beta \right) \right] \left[ Q_{\mathrm{F}_3\left( M \right)}^{F_2}\left( \vec{i}_{2}^{F},\alpha \right) \right] ] [QMF]=[QF1F(kF,γ)][QF2F1(j1F,β)][QF3(M)F2(i2F,α)]]
而对于RPY角(滚转角Roll,仰俯角Pitch,偏航角Yaw)而言,其旋转变换描述为:绕固定坐标系的基矢量
i
⃗
F
\vec{i}^{F}
iF回转
α
\alpha
α,得到新标架
{
F
1
:
(
i
⃗
1
F
,
j
⃗
1
F
,
k
⃗
1
F
)
}
\left\{ F_1:\left( \vec{i}_{1}^{F},\vec{j}_{1}^{F},\vec{k}_{1}^{F} \right) \right\}
{F1:(i1F,j1F,k1F)};再绕固定坐标系的基矢量
j
⃗
F
\vec{j}^{F}
jF回转
β
\beta
β,得到新标架
{
F
2
:
(
i
⃗
2
F
,
j
⃗
2
F
,
k
⃗
2
F
)
}
\left\{ F_2:\left( \vec{i}_{2}^{F},\vec{j}_{2}^{F},\vec{k}_{2}^{F} \right) \right\}
{F2:(i2F,j2F,k2F)};最后绕固定坐标系的基矢量
k
⃗
F
\vec{k}^{F}
kF回转
γ
\gamma
γ,得到新标架
{
F
3
:
(
i
⃗
3
F
,
j
⃗
3
F
,
k
⃗
3
F
)
}
\left\{ F_3:\left( \vec{i}_{3}^{F},\vec{j}_{3}^{F},\vec{k}_{3}^{F} \right) \right\}
{F3:(i3F,j3F,k3F)}为最终的变换姿态(固定坐标系下连续转动,左乘)。因此对于多次连续转动而言有:
[
Q
M
F
]
=
[
Q
F
3
(
M
)
F
2
(
k
⃗
F
,
γ
)
]
[
Q
F
2
F
1
(
j
⃗
F
,
β
)
]
[
Q
F
1
F
(
i
⃗
F
,
α
)
]
\left[ Q_{\mathrm{M}}^{F} \right] =\left[ Q_{\mathrm{F}_3\left( M \right)}^{F_2}\left( \vec{k}^F,\gamma \right) \right] \left[ Q_{\mathrm{F}_2}^{F_1}\left( \vec{j}^F,\beta \right) \right] \left[ Q_{\mathrm{F}_1}^{F}\left( \vec{i}^F,\alpha \right) \right]
[QMF]=[QF3(M)F2(kF,γ)][QF2F1(jF,β)][QF1F(iF,α)]
进而求解出其转换矩阵为:
[
Q
M
F
]
=
[
cos
β
cos
γ
−
cos
β
sin
γ
sin
β
sin
α
sin
β
cos
γ
+
cos
α
sin
γ
−
sin
α
sin
β
sin
γ
+
cos
α
cos
γ
−
sin
α
cos
β
−
cos
α
sin
β
cos
γ
+
sin
α
sin
γ
cos
α
sin
β
cos
γ
+
sin
α
cos
γ
cos
α
cos
β
]
\left[ Q_{\mathrm{M}}^{F} \right] =\left[ \begin{matrix} \cos \beta \cos \gamma& -\cos \beta \sin \gamma& \sin \beta\\ \sin \alpha \sin \beta \cos \gamma +\cos \alpha \sin \gamma& -\sin \alpha \sin \beta \sin \gamma +\cos \alpha \cos \gamma& -\sin \alpha \cos \beta\\ -\cos \alpha \sin \beta \cos \gamma +\sin \alpha \sin \gamma& \cos \alpha \sin \beta \cos \gamma +\sin \alpha \cos \gamma& \cos \alpha \cos \beta\\ \end{matrix} \right]
[QMF]=
cosβcosγsinαsinβcosγ+cosαsinγ−cosαsinβcosγ+sinαsinγ−cosβsinγ−sinαsinβsinγ+cosαcosγcosαsinβcosγ+sinαcosγsinβ−sinαcosβcosαcosβ
同理,若已知转换矩阵:
[
Q
B
A
]
=
[
q
11
q
12
q
13
q
21
q
22
q
23
q
31
q
32
q
33
]
\left[ Q_{\mathrm{B}}^{A} \right] =\left[ \begin{matrix} q_{11}& q_{12}& q_{13}\\ q_{21}& q_{22}& q_{23}\\ q_{31}& q_{32}& q_{33}\\ \end{matrix} \right]
[QBA]=
q11q21q31q12q22q32q13q23q33
, 则有:
[
α
β
γ
]
=
[
a
r
c
tan
(
−
q
23
q
33
)
a
r
c
sin
(
q
13
)
a
r
c
tan
(
−
q
12
q
11
)
]
\left[ \begin{array}{c} \alpha\\ \beta\\ \gamma\\ \end{array} \right] =\left[ \begin{array}{c} \mathrm{arc}\tan \left( -\frac{q_{23}}{q_{33}} \right)\\ \mathrm{arc}\sin \left( q_{13} \right)\\ \mathrm{arc}\tan \left( -\frac{q_{12}}{q_{11}} \right)\\ \end{array} \right]
αβγ
=
arctan(−q33q23)arcsin(q13)arctan(−q11q12)
3.5.1 欧拉角与四元数的转换
- 欧拉角转换为四元数:
q ⃗ = [ cos α 2 cos β 2 cos γ 2 + sin α 2 sin β 2 sin γ 2 sin α 2 cos β 2 cos γ 2 − cos α 2 sin β 2 sin γ 2 cos α 2 sin β 2 cos γ 2 + sin α 2 cos β 2 sin γ 2 cos α 2 cos β 2 sin γ 2 − sin α 2 sin β 2 cos γ 2 ] \vec{q}=\left[ \begin{array}{c} \cos \frac{\alpha}{2}\cos \frac{\beta}{2}\cos \frac{\gamma}{2}+\sin \frac{\alpha}{2}\sin \frac{\beta}{2}\sin \frac{\gamma}{2}\\ \sin \frac{\alpha}{2}\cos \frac{\beta}{2}\cos \frac{\gamma}{2}-\cos \frac{\alpha}{2}\sin \frac{\beta}{2}\sin \frac{\gamma}{2}\\ \cos \frac{\alpha}{2}\sin \frac{\beta}{2}\cos \frac{\gamma}{2}+\sin \frac{\alpha}{2}\cos \frac{\beta}{2}\sin \frac{\gamma}{2}\\ \cos \frac{\alpha}{2}\cos \frac{\beta}{2}\sin \frac{\gamma}{2}-\sin \frac{\alpha}{2}\sin \frac{\beta}{2}\cos \frac{\gamma}{2}\\ \end{array} \right] q= cos2αcos2βcos2γ+sin2αsin2βsin2γsin2αcos2βcos2γ−cos2αsin2βsin2γcos2αsin2βcos2γ+sin2αcos2βsin2γcos2αcos2βsin2γ−sin2αsin2βcos2γ - 四元数转换为欧拉角:
[ α β γ ] = [ a r c tan 2 ( q 1 q 2 + q 3 q 4 ) 1 − 2 ( q 1 2 + q 2 2 ) a r c sin ( 2 ( q 1 q 3 − q 2 q 4 ) ) a r c tan 2 ( q 1 q 2 + q 3 q 4 ) 1 − 2 ( q 1 2 + q 2 2 ) ] \left[ \begin{array}{c} \alpha\\ \beta\\ \gamma\\ \end{array} \right] =\left[ \begin{array}{c} \mathrm{arc}\tan \frac{2\left( q_1q_2+q_3q_4 \right)}{1-2\left( {q_1}^2+{q_2}^2 \right)}\\ \mathrm{arc}\sin \left( 2\left( q_1q_3-q_2q_4 \right) \right)\\ \mathrm{arc}\tan \frac{2\left( q_1q_2+q_3q_4 \right)}{1-2\left( {q_1}^2+{q_2}^2 \right)}\\ \end{array} \right] αβγ = arctan1−2(q12+q22)2(q1q2+q3q4)arcsin(2(q1q3−q2q4))arctan1−2(q12+q22)2(q1q2+q3q4)
3.5.2 欧拉角与轴角的转换
- 欧拉角转换为轴角:
θ = a r c cos ( cos β cos γ − sin α sin β sin γ + cos α ( cos γ + cos β ) − 1 2 ) v ⃗ F = 1 2 sin θ [ cos α sin β cos γ + sin α ( cos γ + cos β ) sin β ( 1 − cos α cos γ ) − sin α sin γ sin α sin β cos γ + sin γ ( cos α + cos β ) ] \begin{split} \theta &=\mathrm{arc}\cos \left( \frac{\cos \beta \cos \gamma -\sin \alpha \sin \beta \sin \gamma +\cos \alpha \left( \cos \gamma +\cos \beta \right) -1}{2} \right) \\ \vec{v}^F&=\frac{1}{2\sin \theta}\left[ \begin{array}{c} \cos \alpha \sin \beta \cos \gamma +\sin \alpha \left( \cos \gamma +\cos \beta \right)\\ \sin \beta \left( 1-\cos \alpha \cos \gamma \right) -\sin \alpha \sin \gamma\\ \sin \alpha \sin \beta \cos \gamma +\sin \gamma \left( \cos \alpha +\cos \beta \right)\\ \end{array} \right] \end{split} θvF=arccos(2cosβcosγ−sinαsinβsinγ+cosα(cosγ+cosβ)−1)=2sinθ1 cosαsinβcosγ+sinα(cosγ+cosβ)sinβ(1−cosαcosγ)−sinαsinγsinαsinβcosγ+sinγ(cosα+cosβ) - 轴角转换为欧拉角:
[ α β γ ] = [ a r c tan ( − q 23 q 33 ) a r c sin ( q 13 ) a r c tan ( − q 12 q 11 ) ] = [ a r c tan v 1 A sin θ − v 2 A v 3 A ( 1 − cos θ ) ( v 3 A ) 2 ( 1 − cos θ ) + cos θ a r c sin ( v 1 A v 3 A ( 1 − cos θ ) + v 2 A sin θ ) a r c tan v 3 A sin θ − v 1 A v 2 A ( 1 − cos θ ) ( v 1 A ) 2 ( 1 − cos θ ) + cos θ ] \left[ \begin{array}{c} \alpha\\ \beta\\ \gamma\\ \end{array} \right] =\left[ \begin{array}{c} \mathrm{arc}\tan \left( -\frac{q_{23}}{q_{33}} \right)\\ \mathrm{arc}\sin \left( q_{13} \right)\\ \mathrm{arc}\tan \left( -\frac{q_{12}}{q_{11}} \right)\\ \end{array} \right] =\left[ \begin{array}{c} \mathrm{arc}\tan \frac{v_{1}^{A}\sin \theta -v_{2}^{A}v_{3}^{A}\left( 1-\cos \theta \right)}{\left( v_{3}^{A} \right) ^2\left( 1-\cos \theta \right) +\cos \theta}\\ \mathrm{arc}\sin \left( v_{1}^{A}v_{3}^{A}\left( 1-\cos \theta \right) +v_{2}^{A}\sin \theta \right)\\ \mathrm{arc}\tan \frac{v_{3}^{A}\sin \theta -v_{1}^{A}v_{2}^{A}\left( 1-\cos \theta \right)}{\left( v_{1}^{A} \right) ^2\left( 1-\cos \theta \right) +\cos \theta}\\ \end{array} \right] αβγ = arctan(−q33q23)arcsin(q13)arctan(−q11q12) = arctan(v3A)2(1−cosθ)+cosθv1Asinθ−v2Av3A(1−cosθ)arcsin(v1Av3A(1−cosθ)+v2Asinθ)arctan(v1A)2(1−cosθ)+cosθv3Asinθ−v1Av2A(1−cosθ)
3.6 连续转动
- 固定坐标系下运动基矢量连续转动:俗称为左乘)即每一次转动后,新的转动轴与刚体上的矢量在固定坐标系下重新定义,用数学公式表达为:
r ⃗ ′ F = [ Q M F ] r ⃗ F = [ Q M F n − 1 ] [ Q F n − 1 F n − 2 ] ⋯ [ Q F 1 F ] r ⃗ F = [ Q M F n − 1 ] [ Q F n − 1 F n − 2 ] ⋯ [ Q F 2 F 1 ] r ⃗ 1 F = [ Q M F n − 1 ] [ Q F n − 1 F n − 2 ] ⋯ [ Q F 3 F 2 ] r ⃗ 2 F = e θ n v ⃗ n F ⋯ e θ 2 v ⃗ 2 F e θ 1 v ⃗ 1 F r ⃗ F \begin{split} {\vec{r}^{\prime}}^F=\left[ Q_{\mathrm{M}}^{F} \right] \vec{r}^F=\left[ Q_{\mathrm{M}}^{F_{n-1}} \right] \left[ Q_{F_{n-1}}^{F_{n-2}} \right] \cdots \left[ Q_{F_1}^{F} \right] \vec{r}^F&=\left[ Q_{\mathrm{M}}^{F_{n-1}} \right] \left[ Q_{F_{n-1}}^{F_{n-2}} \right] \cdots \left[ Q_{F_2}^{F_1} \right] {\vec{r}_1}^F=\left[ Q_{\mathrm{M}}^{F_{n-1}} \right] \left[ Q_{F_{n-1}}^{F_{n-2}} \right] \cdots \left[ Q_{F_3}^{F_2} \right]{\vec{r}_2}^F \\ &=e^{\theta _{\mathrm{n}}{\vec{v}_{\mathrm{n}}}^F}\cdots e^{\theta _2{\vec{v}_2}^F}e^{\theta _1{\vec{v}_1}^F}\vec{r}^F \end{split} r′F=[QMF]rF=[QMFn−1][QFn−1Fn−2]⋯[QF1F]rF=[QMFn−1][QFn−1Fn−2]⋯[QF2F1]r1F=[QMFn−1][QFn−1Fn−2]⋯[QF3F2]r2F=eθnvnF⋯eθ2v2Feθ1v1FrF - 运动坐标系下运动基矢量连续转动:(俗称为右乘)即每一次转动后,新的转动轴与刚体上的矢量在上一次的运动坐标系下重新定义,用数学公式表达为:
r ⃗ ′ F = [ Q M F ] r ⃗ F = [ Q F 1 F ] [ Q F 2 F 1 ] ⋯ [ Q M F n − 1 ] r ⃗ F = e θ 1 v ⃗ 1 F e θ 2 v ⃗ 2 F 1 ⋯ e θ n v ⃗ n F n − 1 r ⃗ F {\vec{r}^{\prime}}^F=\left[ Q_{\mathrm{M}}^{F} \right] \vec{r}^F=\left[ Q_{F_1}^{F} \right] \left[ Q_{F_2}^{F_1} \right] \cdots \left[ Q_{\mathrm{M}}^{F_{n-1}} \right] \vec{r}^F=e^{\theta _1{\vec{v}_1}^F}e^{\theta _2{\vec{v}_2}^{F_1}}\cdots e^{\theta _{\mathrm{n}}{\vec{v}_{\mathrm{n}}}^{F_{n-1}}}\vec{r}^F r′F=[QMF]rF=[QF1F][QF2F1]⋯[QMFn−1]rF=eθ1v1Feθ2v2F1⋯eθnvnFn−1rF
绕两条不同轴进行转动的转换矩阵乘积不可交换。
3.7 齐次矩阵的表达
3.1~3.6 只考虑了坐标系姿态的表达,专注于如何求解/表达
[
Q
M
F
]
\left[ Q_{\mathrm{M}}^{F} \right]
[QMF], 而对于更一般的情况:(忽略原点重合)
R
⃗
P
F
=
[
Q
M
F
]
R
⃗
P
M
+
R
⃗
M
F
\vec{R}_{\mathrm{P}}^{F}=\left[ Q_{\mathrm{M}}^{F} \right] \vec{R}_{\mathrm{P}}^{M}+\vec{R}_{\mathrm{M}}^{F}
RPF=[QMF]RPM+RMF
引入齐次矩阵Homogeneous Transformation Matrix
:
[
T
M
F
]
\left[ T_{\mathrm{M}}^{F} \right]
[TMF]
R
⃗
P
F
=
[
Q
M
F
]
R
⃗
P
M
+
R
⃗
M
F
⇒
[
R
⃗
P
F
1
]
=
[
[
Q
F
M
]
R
⃗
M
F
0
1
×
3
1
]
4
×
4
[
R
⃗
P
M
1
]
\vec{R}_{\mathrm{P}}^{F}=\left[ Q_{\mathrm{M}}^{F} \right] \vec{R}_{\mathrm{P}}^{M}+\vec{R}_{\mathrm{M}}^{F}\Rightarrow \left[ \begin{array}{c} \vec{R}_{\mathrm{P}}^{F}\\ 1\\ \end{array} \right] =\left[ \begin{matrix} \left[ Q_{\mathrm{F}}^{M} \right]& \vec{R}_{\mathrm{M}}^{F}\\ 0_{1\times 3}& 1\\ \end{matrix} \right] _{4\times 4}\left[ \begin{array}{c} \vec{R}_{\mathrm{P}}^{M}\\ 1\\ \end{array} \right]
RPF=[QMF]RPM+RMF⇒[RPF1]=[[QFM]01×3RMF1]4×4[RPM1]
⇒
[
T
M
F
]
=
[
[
Q
M
F
]
R
⃗
M
F
0
1
]
\Rightarrow \left[ T_{\mathrm{M}}^{F} \right] =\left[ \begin{matrix} \left[ Q_{\mathrm{M}}^{F} \right]& \vec{R}_{\mathrm{M}}^{F}\\ 0& 1\\ \end{matrix} \right]
⇒[TMF]=[[QMF]0RMF1]
令:
[
R
⃗
P
F
]
=
[
R
⃗
P
F
1
]
∈
R
4
\left[ \vec{R}_{\mathrm{P}}^{F} \right] =\left[ \begin{array}{c} \vec{R}_{\mathrm{P}}^{F}\\ 1\\ \end{array} \right] \in \mathbb{R} ^4
[RPF]=[RPF1]∈R4,则有:
[
R
⃗
P
F
]
=
[
T
M
F
]
[
R
⃗
P
M
]
\left[ \vec{R}_{\mathrm{P}}^{F} \right] =\left[ T_{\mathrm{M}}^{F} \right] \left[ \vec{R}_{\mathrm{P}}^{M} \right]
[RPF]=[TMF][RPM]
对于向量 R ⃗ P 1 P 2 F \vec{R}_{\mathrm{P}_1\mathrm{P}_2}^{F} RP1P2F 而言,则有:
[ R ⃗ P 1 P 2 F ] = [ R ⃗ P 2 F − R ⃗ P 1 F ] = [ R ⃗ P 2 F 1 ] − [ R ⃗ P 1 F 1 ] = [ R ⃗ P 2 F − R ⃗ P 1 F 0 ] = [ R ⃗ P 1 P 2 F 0 ] \left[ \vec{R}_{\mathrm{P}_1\mathrm{P}_2}^{F} \right] =\left[ \vec{R}_{\mathrm{P}_2}^{F}-\vec{R}_{\mathrm{P}_1}^{F} \right] =\left[ \begin{array}{c} \vec{R}_{\mathrm{P}_2}^{F}\\ 1\\ \end{array} \right] -\left[ \begin{array}{c} \vec{R}_{\mathrm{P}_1}^{F}\\ 1\\ \end{array} \right] =\left[ \begin{array}{c} \vec{R}_{\mathrm{P}_2}^{F}-\vec{R}_{\mathrm{P}_1}^{F}\\ 0\\ \end{array} \right] =\left[ \begin{array}{c} \vec{R}_{\mathrm{P}_1\mathrm{P}_2}^{F}\\ 0\\ \end{array} \right] [RP1P2F]=[RP2F−RP1F]=[RP2F1]−[RP1F1]=[RP2F−RP1F0]=[RP1P2F0]
3.8 点与向量在不同坐标系下的表达
对于固定坐标系下同一点/向量,在不同坐标系
{
A
}
,
{
B
}
\left\{ A \right\} ,\left\{ B \right\}
{A},{B}下进行表达,存在如下转换关系:
R
⃗
V
e
c
t
o
r
A
=
[
Q
B
A
]
R
⃗
V
e
c
t
o
r
B
\vec{R}_{\mathrm{Vector}}^{A}=\left[ Q_{\mathrm{B}}^{A} \right] \vec{R}_{\mathrm{Vector}}^{B}
RVectorA=[QBA]RVectorB
R
⃗
P
A
=
[
Q
B
A
]
R
⃗
P
B
+
R
⃗
B
A
\vec{R}_{\mathrm{P}}^{A}=\left[ Q_{\mathrm{B}}^{A} \right] \vec{R}_{\mathrm{P}}^{B}+\vec{R}_{\mathrm{B}}^{A}
RPA=[QBA]RPB+RBA
1
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
9