本博客是《预积分总结与公式推导》一文的部分学习笔记,记录了一些李群李代数概念,IMU模型与预积分定义等知识。原文写的很好,逻辑清晰,推导详细。但是推导细节过多,不容易把控行文主线。本文提取并记录了该文的脉络核心,略去了细节,便于日后参考查找。本博客只是原文前半段的内容,后续会继续更新。
ω ∧ = [ ω 1 ω 2 ω 3 ] = [ 0 − ω 3 ω 2 ω 3 0 − ω 1 − ω 2 ω 1 0 ] = W \boldsymbol { \omega } ^ { \wedge } = \left[ \begin{array} { c } { \omega _ { 1 } } \\ { \omega _ { 2 } } \\ { \omega _ { 3 } } \end{array} \right] = \left[ \begin{array} { c c c } { 0 } & { - \omega _ { 3 } } & { \omega _ { 2 } } \\ { \omega _ { 3 } } & { 0 } & { - \omega _ { 1 } } \\ { - \omega _ { 2 } } & { \omega _ { 1 } } & { 0 } \end{array} \right] = \mathbf { W } ω∧=⎣⎡ω1ω2ω3⎦⎤=⎣⎡0ω3−ω2−ω30ω1ω2−ω10⎦⎤=W
W ∨ = [ 0 − ω 3 ω 2 ω 3 0 − ω 1 − ω 2 ω 1 0 ] ∨ = [ ω 1 ω 2 ω 3 ] = ω \mathbf { W } ^ { \vee } = \left[ \begin{array} { c c c } { 0 } & { - \omega _ { 3 } } & { \omega _ { 2 } } \\ { \omega _ { 3 } } & { 0 } & { - \omega _ { 1 } } \\ { - \omega _ { 2 } } & { \omega _ { 1 } } & { 0 } \end{array} \right] ^ { \vee } = \left[ \begin{array} { c } { \omega _ { 1 } } \\ { \omega _ { 2 } } \\ { \omega _ { 3 } } \end{array} \right] = \boldsymbol { \omega } W∨=⎣⎡0ω3−ω2−ω30ω1ω2−ω10⎦⎤∨=⎣⎡ω1ω2ω3⎦⎤=ω
a ∧ ⋅ b = − b ∧ ⋅ a , ∀ a , b ∈ R 3 \mathbf { a } ^ { \wedge } \cdot \mathbf { b } = - \mathbf { b } ^ { \wedge } \cdot \mathbf { a } , \quad \forall \mathbf { a } , \mathbf { b } \in R ^ { 3 } a∧⋅b=−b∧⋅a,∀a,b∈R3
exp ( ϕ ⃗ ∧ ) = I + sin ( ∥ ϕ ‾ ∥ ) ∥ ϕ ⃗ ∥ ϕ ⃗ ∧ + 1 − cos ( ∥ ϕ ‾ ∥ ) ∥ ϕ ⃗ ∥ 2 ( ϕ ⃗ ∧ ) 2 \exp \left( \vec { \phi } ^ { \wedge } \right) = \mathbf { I } + \frac { \sin ( \| \overline { \phi } \| ) } { \| \vec { \phi } \| } \vec { \phi } ^ { \wedge } + \frac { 1 - \cos ( \| \overline { \phi } \| ) } { \| \vec { \phi } \| ^ { 2 } } \left( \vec { \phi } ^ { \wedge } \right) ^ { 2 } exp(ϕ∧)=I+∥ϕ∥sin(∥ϕ∥)ϕ∧+∥ϕ∥21−cos(∥ϕ∥)(ϕ∧)2
exp ( ϕ ⃗ ∧ ) ≈ I + ϕ ⃗ ∧ \exp \left( \vec { \phi } ^ { \wedge } \right) \approx \mathbf { I } + \vec { \phi } ^ { \wedge } exp(ϕ∧)≈I+ϕ∧
log ( R ) = φ ⋅ ( R − R T ) 2 sin ( φ ) \log ( \mathbf { R } ) = \frac { \varphi \cdot \left( \mathbf { R } - \mathbf { R } ^ { T } \right) } { 2 \sin ( \varphi ) } log(R)=2sin(φ)φ⋅(R−RT)
其中 φ = cos − 1 ( tr ( R ) − 1 2 ) \varphi = \cos ^ { - 1 } \left( \frac { \operatorname { tr } ( R ) - 1 } { 2 } \right) φ=cos−1(2tr(R)−1)
Exp : R 3 ∋ ϕ ⃗ → exp ( ϕ ⃗ ∧ ) ∈ S O ( 3 ) Log : S O ( 3 ) ∋ R → log ( R ) ∨ ∈ R 3 \begin{array} { l } { \operatorname { Exp } : R ^ { 3 } \ni \vec { \phi } \rightarrow \exp \left( \vec { \phi } ^ { \wedge } \right) \in S O ( 3 ) } \\ { \operatorname{Log} : \mathrm { SO } ( 3 ) \ni \mathbf { R } \rightarrow \log ( \mathbf { R } ) ^ { \vee } \in R ^ { 3 } } \end{array} Exp:R3∋ϕ→exp(ϕ∧)∈SO(3)Log:SO(3)∋R→log(R)∨∈R3
Exp ( ϕ ⃗ + δ ϕ ⃗ ) ≈ Exp ( ϕ ⃗ ) ⋅ Exp ( J r ( ϕ ⃗ ) ⋅ δ ϕ ⃗ ) Log ( Exp ( ϕ ⃗ ) ⋅ Exp ( δ ϕ ⃗ ) ) = ϕ ⃗ + J r − 1 ( ϕ ⃗ ) ⋅ δ ϕ ⃗ \operatorname { Exp } ( \vec { \phi } + \delta \vec { \phi } ) \approx \operatorname { Exp } ( \vec { \phi } ) \cdot \operatorname { Exp } \left( \mathbf { J } _ { r } ( \vec { \phi } ) \cdot \delta \vec { \phi } \right)\\ \operatorname{Log} ( \operatorname { Exp } ( \vec { \phi } ) \cdot \operatorname { Exp } ( \delta \vec { \phi } ) ) = \vec { \phi } + \mathbf { J } _ { r } ^ { - 1 } ( \vec { \phi } ) \cdot \delta \vec { \phi } Exp(ϕ+δϕ)≈Exp(ϕ)⋅Exp(Jr(ϕ)⋅δϕ)Log(Exp(ϕ)⋅Exp(δϕ))=ϕ+Jr−1(ϕ)⋅δϕ
R ⋅ Exp ( ϕ ⃗ ) ⋅ R T = exp ( R ϕ ⃗ ∧ R T ) = Exp ( R ϕ ⃗ ) \mathbf { R } \cdot \operatorname { Exp } ( \vec { \phi } ) \cdot \mathbf { R } ^ { T } = \exp \left( \mathbf { R } \vec { \phi } ^ { \wedge } \mathbf { R } ^ { T } \right) = \operatorname { Exp } ( \mathbf { R } \vec { \phi } ) R⋅Exp(ϕ)⋅RT=exp(Rϕ∧RT)=Exp(Rϕ)
Exp ( ϕ ⃗ ) ⋅ R = R ⋅ Exp ( R T ϕ ⃗ ) \operatorname { Exp } ( \vec { \phi } ) \cdot \mathbf { R } = \mathbf { R } \cdot \operatorname { Exp } \left( \mathbf { R } ^ { T } \vec { \phi } \right) Exp(ϕ)⋅R=R⋅Exp(RTϕ)
( R ϕ ⃗ ) ∧ = R ϕ ⃗ ∧ R T 矩 阵 乘 法 交 换 形 式 : ϕ ⃗ ∧ R = R ( R T ϕ ⃗ ) ∧ ( \mathbf { R } \vec { \phi } ) ^ { \wedge } = \mathbf { R } \vec { \phi } ^ { \wedge } \mathbf { R } ^ { T }\\ 矩阵乘法交换形式:\vec { \phi } ^ { \wedge } \mathbf { R } =\mathbf { R }( \mathbf { R }^T \vec { \phi } ) ^ { \wedge } (Rϕ)∧=Rϕ∧RT矩阵乘法交换形式:ϕ∧R=R(RTϕ)∧
ω ~ w b b ( t ) = ω w b b ( t ) + b g ( t ) + η g ( t ) \tilde { \boldsymbol { \omega } } _ { w b } ^ { b } ( t ) = \boldsymbol { \omega } _ { w b } ^ { b } ( t ) + \mathbf { b } _ { g } ( t ) + \boldsymbol { \eta } _ { g } ( t ) ω~wbb(t)=ωwbb(t)+bg(t)+ηg(t)
翻译:角速度测量值=角速度真实值+角速度测量偏置+角速度测量噪声
加速度和角速度都是在body坐标系(IMU当前坐标系)下进行测量的,但是由于重力加速度是在world坐标系下表示的,加速度真实值也表示在world坐标系下,所以测量值和真实值之间需要body-world坐标变换。由于只是向量坐标变换,只需要乘上旋转量。
f b ( t ) = R b w T ( a w − g w ) + b a ( t ) + η a ( t ) \mathbf { f } ^ { b } ( t ) = \mathbf { R } _ { b } ^ { w T } \left( \mathbf { a } ^ { w } - \mathbf { g } ^ { w } \right) + \mathbf { b } _ { a } ( t ) + \mathbf { \eta } _ { a } ( t ) fb(t)=RbwT(aw−gw)+ba(t)+ηa(t)
翻译:body坐标系下的加速度测量值=加速度真实值减去重力加速度(world坐标系下),转换到body坐标系下后再加上加速度测量偏置+加速度测量噪声
真实值=测量值-测量偏置-测量噪声
ω w b b ( t ) = ω ~ w b b ( t ) − b g ( t ) − η g ( t ) \boldsymbol { \omega } _ { w b } ^ { b } ( t )=\tilde { \boldsymbol { \omega } } _ { w b } ^ { b } ( t ) - \mathbf { b } _ { g } ( t ) - \boldsymbol { \eta } _ { g } ( t ) ωwbb(t)=ω~wbb(t)−bg(t)−ηg(t)
a w = R b w ( f b ( t ) − b a ( t ) − η a ( t ) ) + g w \mathbf { a } ^ { w } =\mathbf { R } _ { b } ^ { w } \left(\mathbf { f } ^ { b } ( t ) - \mathbf { b } _ { a } ( t ) - \mathbf { \eta } _ { a } ( t )\right)+ \mathbf { g } ^ { w } aw=Rbw(fb(t)−ba(t)−ηa(t))+gw
R ˙ b w = R b w ( ω w b b ) ∧ v ˙ w = a w p ˙ w = v w \dot { \mathbf { R } } _ { b } ^ { w } = \mathbf { R } _ { b } ^ { w } \left( \boldsymbol { \omega } _ { w b } ^ { b } \right) ^ { \wedge } \\ \dot { \mathbf { v } } ^ { w } = \mathbf { a } ^ { w } \\ \dot { \mathbf { p } } ^ { w } = \mathbf { v } ^ { w } R˙bw=Rbw(ωwbb)∧v˙w=awp˙w=vw
翻译:旋转矩阵对时间的导数=旋转矩阵*旋转矩阵对应的李代数;速度对时间导数=角速度;位置对时间的导数=速度
R b ( t + Δ t ) w = R b ( t ) w Exp ( ω w b b ( t ) ⋅ Δ t ) v w ( t + Δ t ) = v w ( t ) + a w ( t ) ⋅ Δ t p w ( t + Δ t ) = p w ( t ) + v w ( t ) ⋅ Δ t + 1 2 a w ( t ) ⋅ Δ t 2 \begin{array} { l } { \mathbf { R } _ { b ( t + \Delta t ) } ^ { w } = \mathbf { R } _ { b ( t ) } ^ { w } \operatorname { Exp } \left( \boldsymbol { \omega } _ { w b } ^ { b } ( t ) \cdot \Delta t \right) } \\ { \mathbf { v } ^ { w } ( t + \Delta t ) = \mathbf { v } ^ { w } ( t ) + \mathbf { a } ^ { w } ( t ) \cdot \Delta t } \\ { \mathbf { p } ^ { w } ( t + \Delta t ) = \mathbf { p } ^ { w } ( t ) + \mathbf { v } ^ { w } ( t ) \cdot \Delta t + \frac { 1 } { 2 } \mathbf { a } ^ { w } ( t ) \cdot \Delta t ^ { 2 } } \end{array} Rb(t+Δt)w=Rb(t)wExp(ωwbb(t)⋅Δt)vw(t+Δt)=vw(t)+aw(t)⋅Δtpw(t+Δt)=pw(t)+vw(t)⋅Δt+21aw(t)⋅Δt2
将测量模型的第二种形式代入测量模型的离散形式:
R ( t + Δ t ) = R ( t ) ⋅ Exp ( ( ω ~ ( t ) − b g ( t ) − η g d ( t ) ) ⋅ Δ t ) v ( t + Δ t ) = v ( t ) + R ( t ) ⋅ ( f ~ ( t ) − b a ( t ) − η a d ( t ) ) ⋅ Δ t + g ⋅ Δ t p ( t + Δ t ) = p ( t ) + v ( t ) ⋅ Δ t + 1 2 g ⋅ Δ t 2 + 1 2 R ( t ) ⋅ ( f ~ ( t ) − b a ( t ) − η a d ( t ) ) ⋅ Δ t 2 \mathbf { R } ( t + \Delta t ) =\mathbf { R } ( t ) \cdot \operatorname { Exp } \left( \left( \tilde { \boldsymbol { \omega } } ( t ) - \mathbf { b } _ { g } ( t ) - \mathbf { \eta } _ { g d } ( t ) \right) \cdot \Delta t \right)\\ \mathbf { v } ( t + \Delta t )=\mathbf { v } ( t ) + \mathbf { R } ( t ) \cdot \left( \tilde { \mathbf { f } } ( t ) - \mathbf { b } _ { a } ( t ) - \mathbf { \eta } _ { a d } ( t ) \right) \cdot \Delta t + \mathbf { g } \cdot \Delta t\\ \mathbf { p } ( t + \Delta t ) = \mathbf { p } ( t ) + \mathbf { v } ( t ) \cdot \Delta t + \frac { 1 } { 2 } \mathbf { g } \cdot \Delta t ^ { 2 } + \frac { 1 } { 2 } \mathbf { R } ( t ) \cdot \left( \tilde { \mathbf { f } } ( t ) - \mathbf { b } _ { a } ( t ) - \mathbf { \eta } _ { a d } ( t ) \right) \cdot \Delta t ^ { 2 } R(t+Δt)=R(t)⋅Exp((ω~(t)−bg(t)−ηgd(t))⋅Δt)v(t+Δt)=v(t)+R(t)⋅(f~(t)−ba(t)−ηad(t))⋅Δt+g⋅Δtp(t+Δt)=p(t)+v(t)⋅Δt+21g⋅Δt2+21R(t)⋅(f~(t)−ba(t)−ηad(t))⋅Δt2
为了简化表示,上式省略了一些上下标,并对一些表示进行了调整。同时,使用离散噪声代替连续噪声,二者关系如下:
Cov ( η g d ( t ) ) = 1 Δ t Cov ( η g ( t ) ) Cov ( η a d ( t ) ) = 1 Δ t Cov ( η a ( t ) ) \operatorname { Cov } \left( \boldsymbol { \eta } _ { \mathrm { gd } } ( t ) \right) = \frac { 1 } { \Delta t } \operatorname { Cov } \left( \boldsymbol { \eta } _ { \mathrm { g } } ( t ) \right)\\ \operatorname { Cov } \left( \boldsymbol { \eta } _ { a d } ( t ) \right) = \frac { 1 } { \Delta t } \operatorname { Cov } \left( \mathbf { \eta } _ { a } ( t ) \right) Cov(ηgd(t))=Δt1Cov(ηg(t))Cov(ηad(t))=Δt1Cov(ηa(t))
用离散时间步 k k k代替连续时间 t t t,并进一步简化符号
R k + 1 = R k ⋅ Exp ( ( ω ~ k − b k g − η k g d ) ⋅ Δ t ) v k + 1 = v k + R k ⋅ ( f ~ k − b k a − η k a d ) ⋅ Δ t + g ⋅ Δ t p k + 1 = p k + v k ⋅ Δ t + 1 2 g ⋅ Δ t 2 + 1 2 R k ⋅ ( f ~ k − b k a − η k a d ) ⋅ Δ t 2 \begin{array} { l } { \mathbf { R } _ { k + 1 } = \mathbf { R } _ { k } \cdot \operatorname { Exp } \left( \left( \tilde { \boldsymbol { \omega } } _ { k } - \mathbf { b } _ { k } ^ { g } - \boldsymbol { \eta } _ { k } ^ { g d } \right) \cdot \Delta t \right) } \\ { \mathbf { v } _ { k + 1 } = \mathbf { v } _ { k } + \mathbf { R } _ { k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } - \mathbf { \eta } _ { k } ^ { a d } \right) \cdot \Delta t + \mathbf { g } \cdot \Delta t } \\ { \mathbf { p } _ { k + 1 } = \mathbf { p } _ { k } + \mathbf { v } _ { k } \cdot \Delta t + \frac { 1 } { 2 } \mathbf { g } \cdot \Delta t ^ { 2 } + \frac { 1 } { 2 } \mathbf { R } _ { k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } - \mathbf { \eta } _ { k } ^ { a d } \right) \cdot \Delta t ^ { 2 } } \end{array} Rk+1=Rk⋅Exp((ω~k−bkg−ηkgd)⋅Δt)vk+1=vk+Rk⋅(f~k−bka−ηkad)⋅Δt+g⋅Δtpk+1=pk+vk⋅Δt+21g⋅Δt2+21Rk⋅(f~k−bka−ηkad)⋅Δt2
由时间区间 [ i , j − 1 ] [i,j-1] [i,j−1]内的IMU测量,偏置和噪声,可以积分得到时间步 j j j的IMU旋转,速度和位移量。
R j = R i ⋅ ∏ k = i j − 1 Exp ( ( ω ~ k − b k g − η k g d ) ⋅ Δ t ) v j = v i + g ⋅ Δ t i j + ∑ k = i j − 1 R k ⋅ ( f ~ k − b k a − η k a d ) ⋅ Δ t p j = p i + ∑ k = i j − 1 [ v k ⋅ Δ t + 1 2 g ⋅ Δ t 2 + 1 2 R k ⋅ ( f ~ k − b k a − η k a d ) ⋅ Δ t 2 ] \mathbf { R } _ { j } = \mathbf { R } _ { i } \cdot \prod _ { k = i } ^ { j - 1 } \operatorname { Exp } \left( \left( \tilde { \boldsymbol { \omega } } _ { k } - \mathbf { b } _ { k } ^ { g } - \mathbf { \eta } _ { k } ^ { g d } \right) \cdot \Delta t \right)\\ \mathbf { v } _ { j } = \mathbf { v } _ { i } + \mathbf { g } \cdot \Delta t _ { i j } + \sum _ { k = i } ^ { j - 1 } \mathbf { R } _ { k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } - \mathbf { \eta } _ { k } ^ { a d } \right) \cdot \Delta t\\ \mathbf { p } _ { j } = \mathbf { p } _ { i } + \sum _ { k = i } ^ { j - 1 } \left[ \mathbf { v } _ { k } \cdot \Delta t + \frac { 1 } { 2 } \mathbf { g } \cdot \Delta t ^ { 2 } + \frac { 1 } { 2 } \mathbf { R } _ { k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } - \mathbf { \eta } _ { k } ^ { a d } \right) \cdot \Delta t ^ { 2 } \right] Rj=Ri⋅k=i∏j−1Exp((ω~k−bkg−ηkgd)⋅Δt)vj=vi+g⋅Δtij+k=i∑j−1Rk⋅(f~k−bka−ηkad)⋅Δtpj=pi+k=i∑j−1[vk⋅Δt+21g⋅Δt2+21Rk⋅(f~k−bka−ηkad)⋅Δt2]
虽然有了区间离散积分,但是起点变化后,需要从头开始积分。因此,我们想得到离散积分的增量形式。预积分定义了下述三个增量:
Δ R i j ≜ R i T R j = ∏ k = i j − 1 Exp ( ( ω ~ k − b k g − η k g d ) ⋅ Δ t ) Δ v i j ≜ R i T ( v j − v i − g ⋅ Δ t i j ) = ∑ k = i j − 1 Δ R i k ⋅ ( f ~ k − b k a − η k a d ) ⋅ Δ t Δ p i j ≜ R i T ( p j − p i − v i ⋅ Δ t i j − 1 2 g ⋅ Δ t i j 2 ) = ∑ k = i j − 1 [ Δ v i k ⋅ Δ t + 1 2 Δ R i k ⋅ ( f ~ k − b k a − η k a d ) ⋅ Δ t 2 ] \Delta \mathbf { R } _ { i j } \triangleq \mathbf { R } _ { i } ^ { T } \mathbf { R } _ { j }= \prod _ { k = i } ^ { j - 1 } \operatorname { Exp } \left( \left( \tilde { \boldsymbol { \omega } } _ { k } - \mathbf { b } _ { k } ^ { g } - \boldsymbol { \eta } _ { k } ^ { g d } \right) \cdot \Delta t \right)\\ \Delta \mathbf { v } _ { i j } \triangleq \mathbf { R } _ { i } ^ { T } \left( \mathbf { v } _ { j } - \mathbf { v } _ { i } - \mathbf { g } \cdot \Delta t _ { i j } \right)=\sum _ { k = i } ^ { j - 1 } \Delta \mathbf { R } _ { i k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } - \mathbf { \eta } _ { k } ^ { a d } \right) \cdot \Delta t\\ \Delta \mathbf { p } _ { i j } \triangleq \mathbf { R } _ { i } ^ { T } \left( \mathbf { p } _ { j } - \mathbf { p } _ { i } - \mathbf { v } _ { i } \cdot \Delta t _ { i j } - \frac { 1 } { 2 } \mathbf { g } \cdot \Delta t _ { i j } ^ { 2 } \right)\\=\sum _ { k = i } ^ { j - 1 } \left[ \Delta \mathbf { v } _ { i k } \cdot \Delta t + \frac { 1 } { 2 } \Delta \mathbf { R } _ { i k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } - \mathbf { \eta } _ { k } ^ { a d } \right) \cdot \Delta t ^ { 2 } \right] ΔRij≜RiTRj=k=i∏j−1Exp((ω~k−bkg−ηkgd)⋅Δt)Δvij≜RiT(vj−vi−g⋅Δtij)=k=i∑j−1ΔRik⋅(f~k−bka−ηkad)⋅ΔtΔpij≜RiT(pj−pi−vi⋅Δtij−21g⋅Δtij2)=k=i∑j−1[Δvik⋅Δt+21ΔRik⋅(f~k−bka−ηkad)⋅Δt2]
旋转增量的定义最好理解,直接左乘 i i i时刻的旋转分量即可。速度和位置增量除了减去 i i i时刻的速度和位移外,还额外减去了一部分重力加速度,这是为了在最终的增量结果中约去重力加速度分量。上式中位移增量的证明不容易直接得到,具体推导过程见参考文献[1]。
Δ R i j ≜ Δ R ~ i j ⋅ Exp ( − δ ϕ ⃗ i j ) \Delta \mathbf { R } _ { i j } \triangleq \Delta \tilde { \mathbf { R } } _ { i j } \cdot \operatorname { Exp } \left( - \delta \vec { \phi } _ { i j } \right) ΔRij≜ΔR~ij⋅Exp(−δϕij)
翻译:旋转预积分 (理想值) = 旋转预积分测量值 “减去“ 旋转预积分噪声值。
其中
Δ R ~ i j = ∏ k = i j − 1 Exp ( ( ω ~ k − b k g ) Δ t ) Exp ( − δ ϕ ⃗ i j ) = ∏ k = i j − 1 Exp ( − Δ R ~ k + 1 j T ⋅ J r ( ( ω ~ k − b k g ) Δ t ) ⋅ η k g d Δ t ) \Delta \tilde { \mathbf { R } } _ { i j } = \prod _ { k = i } ^ { j - 1 } \operatorname { Exp } \left( \left( \tilde { \boldsymbol { \omega } } _ { k } - \mathbf { b } _ { k } ^ { g } \right) \Delta t \right)\\ \operatorname { Exp } \left( - \delta \vec { \phi } _ { i j } \right) = \prod _ { k = i } ^ { j - 1 } \operatorname { Exp } \left( - \Delta \tilde { \mathbf { R } } _ { k + 1 j } ^ { T } \cdot \mathbf { J } _ { r } \left( \left( \tilde { \boldsymbol { \omega } } _ { k } - \mathbf { b } _ { k } ^ { g } \right) \Delta t \right) \cdot \boldsymbol { \eta } _ { k } ^ { g d } \Delta t \right) ΔR~ij=k=i∏j−1Exp((ω~k−bkg)Δt)Exp(−δϕij)=k=i∏j−1Exp(−ΔR~k+1jT⋅Jr((ω~k−bkg)Δt)⋅ηkgdΔt)
翻译:旋转预积分测量值与IMU角速度测量,偏置有关;旋转预积分噪声值与IMU角速度测量值,偏置,噪声,以及旋转预积分测量值有关。
Δ v i j ≜ Δ v ~ i j − δ v i j \Delta \mathbf { v } _ { i j } \triangleq \Delta \tilde { \mathbf { v } } _ { i j } - \delta \mathbf { v } _ { i j } Δvij≜Δv~ij−δvij
翻译:速度预积分 (理想值) = 速度预积分测量值减去速度预积分噪声值。
其中
Δ v ~ i j ≜ ∑ k = i j − 1 [ Δ R ~ i k ⋅ ( f ~ k − b k a ) ⋅ Δ t ] δ v i j ≜ ∑ k = i j − 1 [ Δ R ~ i k η k a d Δ t − Δ R ~ i k ⋅ ( f ~ k − b k a ) ∧ ⋅ δ ϕ ⃗ i k ⋅ Δ t ] \Delta \tilde { \mathbf { v } } _ { i j } \triangleq \sum _ { k = i } ^ { j - 1 } \left[ \Delta \tilde { \mathbf { R } } _ { i k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } \right) \cdot \Delta t \right]\\ \delta \mathbf { v } _ { i j } \triangleq \sum _ { k = i } ^ { j - 1 } \left[ \Delta \tilde { \mathbf { R } } _ { i k } \mathbf { \eta } _ { k } ^ { a d } \Delta t - \Delta \tilde { \mathbf { R } } _ { i k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } \right) ^ { \wedge } \cdot \delta \vec { \phi } _ { i k } \cdot \Delta t \right] Δv~ij≜k=i∑j−1[ΔR~ik⋅(f~k−bka)⋅Δt]δvij≜k=i∑j−1[ΔR~ikηkadΔt−ΔR~ik⋅(f~k−bka)∧⋅δϕik⋅Δt]
翻译:速度预积分测量值与IMU加速度测量,偏置;旋转预积分测量有关;速度预积分噪声值与IMU加速度测量,偏置,噪声;以及旋转预积分测量,噪声有关。
Δ p i j ≜ Δ p ~ i j − δ p i j \Delta \mathbf { p } _ { i j } \triangleq \Delta \tilde { \mathbf { p } } _ { i j } - \delta \mathbf { p } _ { i j } Δpij≜Δp~ij−δpij
翻译:位置预积分 (理想值) = 位置预积分测量值减去位置预积分噪声值。
其中
Δ p ~ i j ≜ ∑ k = i j − 1 [ Δ v ~ i k Δ t + 1 2 Δ R ~ i k ⋅ ( f ~ k − b k a ) Δ t 2 ] δ p i j ≜ ∑ k = i j − 1 [ δ v i k Δ t − 1 2 Δ R ~ i k ⋅ ( f ~ k − b k a ) ∧ δ ϕ ⃗ i k Δ t 2 + 1 2 Δ R ~ i k η k a d Δ t 2 ] \Delta \tilde { \mathbf { p } } _ { i j } \triangleq \sum _ { k = i } ^ { j - 1 } \left[ \Delta \tilde { \mathbf { v } } _ { i k } \Delta t + \frac { 1 } { 2 } \Delta \tilde { \mathbf { R } } _ { i k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } \right) \Delta t ^ { 2 } \right]\\ \delta \mathbf { p } _ { i j } \triangleq \sum _ { k = i } ^ { j - 1 } \left[ \delta \mathbf { v } _ { i k } \Delta t - \frac { 1 } { 2 } \Delta \tilde { \mathbf { R } } _ { i k } \cdot \left( \tilde { \mathbf { f } } _ { k } - \mathbf { b } _ { k } ^ { a } \right) ^ { \wedge } \delta \vec { \phi } _ { i k } \Delta t ^ { 2 } + \frac { 1 } { 2 } \Delta \tilde { \mathbf { R } } _ { i k } \mathbf { \eta } _ { k } ^ { a d } \Delta t ^ { 2 } \right] Δp~ij≜k=i∑j−1[Δv~ikΔt+21ΔR~ik⋅(f~k−bka)Δt2]δpij≜k=i∑j−1[δvikΔt−21ΔR~ik⋅(f~k−bka)∧δϕikΔt2+21ΔR~ikηkadΔt2]
翻译:位置预积分测量值与IMU加速度测量,偏置;旋转和速度预积分测量有关;速度预积分噪声值与IMU加速度测量,偏置,噪声;旋转预积分测量,噪声;速度预积分噪声有关。
邱笑晨,预积分总结与公式推导,2018