Planning-Frenet坐标和Cartesian坐标转换中

本文根据B站up:忠厚老实的老王 的视频《自动驾驶决策规划算法第一章第三节(上) 直角坐标与自然坐标转换》整理,连接:https://www.bilibili.com/video/BV1tQ4y1r7fh?spm_id_from=333.1007.top_right_bar_window_history.content.click

Planning-Frenet坐标和Cartesian坐标转换上
Planning-Frenet坐标和Cartesian坐标转换下

2 坐标转换

在这里插入图片描述

2.1 变量说明:

r h ⃗ \vec{r_h} rh 车的位置矢量 r h ⃗ \vec{r_h} rh 投影的位置矢量
v ⃗ \vec{v} v 车的速度 s ˙ \dot{s} s˙投影的速率
a ⃗ \vec{a} a 车的加速度
κ h \kappa _h κh车的位置矢量在车轨迹上的曲率 κ r \kappa _r κr投影位置矢量在道路几何上的曲率
τ h ⃗ \vec{\tau_h} τh 车的位置矢量在车轨迹上的切线方向的单位向量 τ r ⃗ \vec{\tau_r} τr 投影位置矢量在道路几何上的切线方向的单位向量
n h ⃗ \vec{n_h} nh 车的位置矢量在车轨迹上的法线方向的单位向量 n r ⃗ \vec{n_r} nr 投影位置矢量在道路几何上的法线方向的单位向量

2.2 基础公式

计算所需要的公式,即为上述 ( 1 − 1 ) — ( 1 − 4 ) (1-1)—(1-4) (11)(14)
{ ① r h ˙ ⃗ = ∣ v h ⃗ ∣ τ h ⃗ = v h ⃗ ② r r ˙ ⃗ = s ˙ τ r ⃗ ③ τ h ˙ ⃗ = d τ h ⃗ d t = κ h ∣ v h ⃗ ∣ n h ⃗ ④ n h ˙ ⃗ = − κ h ∣ v h ⃗ ∣ τ h ⃗ ⑤ τ r ˙ ⃗ = κ r s r ˙ n r ⃗ ⑥ n r ˙ ⃗ = − κ r s r ˙ τ r ⃗ ⑦ a ⃗ = ∣ v h ˙ ⃗ ∣ τ h ⃗ + ∣ v h ⃗ ∣ 2 κ h n h ⃗ (2-1) \begin{cases} ①\vec{\dot{r_h}} = |\vec{v_h}| \vec{\tau _h} = \vec{v_h}\\ ②\vec{\dot{r_r}} = \dot{s} \vec{\tau _r} \\ ③\vec{\dot{\tau_h}} = \frac{d \vec{\tau _h}}{dt} = \kappa_h |\vec{v_h}| \vec{n_h} \\ ④\vec{\dot{n_h}} = - \kappa_h |\vec{v_h}| \vec{\tau_h} \\ ⑤\vec{\dot{\tau_r}} = \kappa_r \dot{s_r} \vec{n_r} \\ ⑥\vec{\dot{n_r}} = - \kappa_r \dot{s_r} \vec{\tau_r} \\ ⑦\vec{a} = |\vec{\dot{v_h}}| \vec{\tau_h} + |\vec{v_h}|^2 \kappa_h \vec{n_h} \end{cases} \tag{2-1} rh˙ =vh τh =vh rr˙ =s˙τr τh˙ =dtdτh =κhvh nh nh˙ =κhvh τh τr˙ =κrsr˙nr nr˙ =κrsr˙τr a =vh˙ τh +vh 2κhnh (2-1)

2.3 C a r t e s i a n Cartesian Cartesian坐标系转换到 F r e n e t Frenet Frenet坐标系

已知 r h ⃗ , v h ⃗ , a h ⃗ , κ h , τ h ⃗ , n h ⃗ \vec{r_h}, \vec{v_h}, \vec{a_h}, \kappa_h, \vec{\tau_h}, \vec{n_h} rh ,vh ,ah ,κh,τh ,nh F r e n e t Frenet Frenet坐标系起点 ( x 0 , y 0 ) (x_0,y_0) (x0,y0),求 s , s ˙ , s ¨ , l , l ′ , l ′ ′ , l ˙ , l ¨ s, \dot{s}, \ddot{s}, l, l^{\prime}, l^{\prime \prime}, \dot{l}, \ddot{l} s,s˙,s¨,l,l,l,l˙,l¨

2.3.1 计算ADCFrenet坐标系中的投影点

需要根据车辆坐标 ( x h , y h ) (x_h, y_h) (xh,yh),在参考路径上找到 x r , y r , θ r , κ r x_r, y_r, \theta_r, \kappa_r xr,yr,θr,κr。首先找到距离车辆坐标 ( x h , y h ) (x_h, y_h) (xh,yh)最近的点,称为匹配点 x m , y m , θ m , κ m x_m, y_m, \theta_m, \kappa_m xm,ym,θm,κm

在这里插入图片描述

根据匹配点和ADC坐标可以得到:

{ d ⃗ = ( x h − x m , y h − y m ) τ m ⃗ = ( c o s θ m , s i n θ m ) (2-2) \begin{cases} \vec{d} = (x_h - x_m, y_h - y_m) \\ \vec{\tau_m} = (cos{\theta_m}, sin{\theta_m}) \tag{2-2} \end{cases} {d =(xhxm,yhym)τm =(cosθm,sinθm)(2-2)

将车辆坐标在曲线投影点处的 t r ⃗ \vec{t_r} tr 近似为 τ ⃗ \vec{\tau} τ ,可以得到:

r r ⃗ = ( x r y r ) = r m ⃗ + ( d ⃗ ⋅ τ m ⃗ ) τ m ⃗ = ( x m y m ) + ( d ⃗ ⋅ τ m ⃗ ) ( c o s θ m s i n θ m ) (2-3) \vec{r_r} = \begin{pmatrix} x_r \\ y_r \end{pmatrix} = \vec{r_m} + (\vec{d} \cdot \vec{\tau_m}) \vec{\tau_m} = \begin{pmatrix} x_m \\ y_m \end{pmatrix} + (\vec{d} \cdot \vec{\tau_m}) \begin{pmatrix} cos{\theta_m} \\ sin{\theta_m} \end{pmatrix} \tag{2-3} rr =(xryr)=rm +(d τm )τm =(xmym)+(d τm )(cosθmsinθm)(2-3)

其中的向量点乘可以转化为以下形式:
d ⃗ ⋅ τ m ⃗ = ∣ d ⃗ ∣ ∣ τ m ⃗ ∣ c o s < θ h , θ r > = ∣ d ⃗ ∣ c o s ( θ h − θ r ) (2-4) \vec{d} \cdot \vec{\tau_m} = |\vec{d}| |\vec{\tau_m}| cos<\theta_h, \theta_r> = |\vec{d}| cos(\theta_h - \theta_r) \tag{2-4} d τm =d τm cos<θh,θr>=d cos(θhθr)(2-4)

为了后续计算表达简明,车辆在 F r e n e t Frenet Frenet坐标系下的投影点在 c a r t e s i a n cartesian cartesian坐标系下的坐标,记为 x r , y r , θ r , κ r x_r, y_r, \theta_r, \kappa_r xr,yr,θr,κr,因此可以得到: r r ⃗ = ( x r , y r ) , τ r ⃗ = ( c o s θ r , s i n θ r ) , n r ⃗ = ( − s i n θ r , c o s θ r ) \vec{r_r} = (x_r,y_r), \vec{\tau_r} = (cos \theta_r, sin \theta_r), \vec{n_r} = (-sin \theta_r, cos \theta_r) rr =(xr,yr),τr =(cosθr,sinθr),nr =(sinθr,cosθr)

利用向量三角形和微积分求出 s , s ˙ , s ¨ , l , l ′ , l ′ ′ , l ˙ , l ¨ s, \dot{s}, \ddot{s}, l, l^{\prime}, l^{\prime \prime}, \dot{l}, \ddot{l} s,s˙,s¨,l,l,l,l˙,l¨

核心公式 r r ⃗ + l × n r ⃗ = r h ⃗ \vec{r_r} + l \times \vec{n_r} = \vec{r_h} rr +l×nr =rh

2.3.2 计算 l l l

公式 r r ⃗ + l n r ⃗ = r h ⃗ \vec{r_r} + l \vec{n_r} = \vec{r_h} rr +lnr =rh 两边点乘 n r ⃗ \vec{n_r} nr 可得:
l = ( r h ⃗ − r r ⃗ ) ⋅ n r ⃗ (2-5) l = (\vec{r_h} - \vec{r_r}) \cdot \vec{n_r} \tag{2-5} l=(rh rr )nr (2-5)
写成三角函数形式:
l = s i g n ( ( y h − y r ) c o s ( θ r ) − ( x h − x r ) s i n ( θ r ) ) ( y h − y r ) 2 − ( x h − x r ) 2 (2-6) l = sign((y_h - y_r) cos(\theta_r) - (x_h - x_r)sin(\theta_r)) \sqrt{(y_h - y_r)^2 - (x_h - x_r)^2} \tag{2-6} l=sign((yhyr)cos(θr)(xhxr)sin(θr))(yhyr)2(xhxr)2 (2-6)

3.3.3 计算 s ˙ ( s r ˙ ) \dot{s}(\dot{s_r}) s˙(sr˙):

公式 r r ⃗ + l × n r ⃗ = r h ⃗ \vec{r_r} + l \times \vec{n_r} = \vec{r_h} rr +l×nr =rh 两边同时对时间求导得:
r r ˙ ⃗ + l n r ˙ ⃗ + l ˙ n r ⃗ = r h ˙ ⃗ (2-7) \vec{\dot{r_r}} + l \vec{\dot{n_r}} + \dot{l} \vec{n_r} = \vec{\dot{r_h}} \tag{2-7} rr˙ +lnr˙ +l˙nr =rh˙ (2-7)
利用基础公式①②⑤⑥得:
s r ˙ τ r ⃗ + l ( − κ r s r ˙ τ r ⃗ ) + l ˙ n r ⃗ = v h ⃗ (2-8) \dot{s_r} \vec{\tau _r} + l (- \kappa_r \dot{s_r} \vec{\tau_r}) + \dot{l} \vec{n_r} = \vec{v_h} \tag{2-8} sr˙τr +l(κrsr˙τr )+l˙nr =vh (2-8)
两边同时点乘 τ r ⃗ \vec{\tau_r} τr 得: s r ˙ + l ( − κ r s r ˙ ) = v h ⃗ ⋅ τ r ⃗ \dot{s_r}+l(-\kappa_r \dot{s_r}) = \vec{v_h} \cdot \vec{\tau_r} sr˙+l(κrsr˙)=vh τr ,得到:
s r ˙ = v h ⃗ ⋅ τ r ⃗ 1 − κ r l (2-9) \dot{s_r} = \frac{\vec{v_h} \cdot \vec{\tau_r}}{1-\kappa_r l} \tag{2-9} sr˙=1κrlvh τr (2-9)
转化为三角函数形式:
v e c v h ⋅ τ r ⃗ = ∣ v h ⃗ ∣ τ h ⃗ ⋅ τ r ⃗ = ∣ v h ⃗ ∣ ∣ τ h ⃗ ∣ ∣ τ r ⃗ ∣ c o s < τ h ⃗ , τ r ⃗ > = ∣ v h ⃗ ∣ c o s ( θ h − θ r ) (2-10) vec{v_h} \cdot \vec{\tau_r} = |\vec{v_h}| \vec{\tau_h} \cdot \vec{\tau_r} = |\vec{v_h}| |\vec{\tau_h}| |\vec{\tau_r}| cos<\vec{\tau_h}, \vec{\tau_r}> = |\vec{v_h}| cos(\theta_h - \theta_r) \tag{2-10} vecvhτr =vh τh τr =vh τh τr cos<τh ,τr >=vh cos(θhθr)(2-10)
其中 ∣ τ h ⃗ ∣ = ∣ τ r ⃗ ∣ = 1 |\vec{\tau_h}| = |\vec{\tau_r}| = 1 τh =τr =1

因此,公式 ( 2 − 9 ) (2-9) (29)可以等价写为:
s r ˙ = ∣ v h ⃗ ∣ c o s ( θ h − θ r ) 1 − κ r l (2-11) \dot{s_r} = \frac{|\vec{v_h}| cos(\theta_h - \theta_r)}{1-\kappa_r l} \tag{2-11} sr˙=1κrlvh cos(θhθr)(2-11)

3.3.4 计算 l ˙ \dot{l} l˙

由公式 ( 2 − 8 ) (2-8) (28) s r ˙ τ r ⃗ + l ( − κ r s r ˙ τ r ⃗ ) + l ˙ n r ⃗ = v h ⃗ \dot{s_r} \vec{\tau _r} + l (- \kappa_r \dot{s_r} \vec{\tau_r}) + \dot{l} \vec{n_r} = \vec{v_h} sr˙τr +l(κrsr˙τr )+l˙nr =vh ,两边同时点乘 n r ⃗ \vec{n_r} nr 得到:
l ˙ = v h ⃗ ⋅ n r ⃗ (2-12) \dot{l} = \vec{v_h} \cdot \vec{n_r} \tag{2-12} l˙=vh nr (2-12)
写成三角函数形式有:
l ˙ = v h ⃗ ⋅ n r ⃗ = ∣ v h ⃗ ∣ ∣ n r ⃗ ∣ c o s < v h ⃗ , n r ⃗ > = ∣ v h ⃗ ∣ c o s ( θ h − ( θ r + π 2 ) ) = ∣ v h ⃗ ∣ s i n ( θ h − θ r ) (2-13) \dot{l} = \vec{v_h} \cdot \vec{n_r} = |\vec{v_h}| |\vec{n_r}| cos<\vec{v_h}, \vec{n_r}> = |\vec{v_h}| cos(\theta_h - (\theta_r + \frac{\pi}{2})) = |\vec{v_h}| sin(\theta_h - \theta_r) \tag{2-13} l˙=vh nr =vh nr cos<vh ,nr >=vh cos(θh(θr+2π))=vh sin(θhθr)(2-13)

3.3.5计算 l ′ l^{\prime} l

l ′ = d l d s = d l d t d s r d t = l ˙ s r ˙ = ( 1 − κ r l ) v h ⃗ ⋅ n r ⃗ v h ⃗ ⋅ τ r ⃗ (2-14) l^{\prime} = \frac{dl}{ds}=\frac{\frac{dl}{dt}}{\frac{ds_r}{dt}}=\frac{\dot{l}}{\dot{s_r}} = (1-\kappa_r l) \frac{\vec{v_h} \cdot \vec{n_r}}{\vec{v_h} \cdot \vec{\tau_r}} \tag{2-14} l=dsdl=dtdsrdtdl=sr˙l˙=(1κrl)vh τr vh nr (2-14)

写成三角函数形式:
l ′ = ( 1 − κ r l ) v h ⃗ ⋅ n r ⃗ v h ⃗ ⋅ τ r ⃗ = ( 1 − κ r l ) t a n ( θ h − θ r ) (2-14) l^{\prime} = (1-\kappa_r l) \frac{\vec{v_h} \cdot \vec{n_r}}{\vec{v_h} \cdot \vec{\tau_r}} = (1-\kappa_r l) tan(\theta_h - \theta_r) \tag{2-14} l=(1κrl)vh τr vh nr =(1κrl)tan(θhθr)(2-14)

3.3.6 计算 s ¨ ( s r ¨ ) \ddot{s}(\ddot{s_r}) s¨(sr¨)

s r ¨ = d ( v h ⃗ ⋅ τ r ⃗ 1 − κ r l ) d t = 1 ( 1 − κ r l ) 2 ( d ( v h ⃗ ⋅ τ r ⃗ ) d t ( 1 − κ r l ) − ( v h ⃗ ⋅ τ r ⃗ ) ( − κ r ˙ l − κ r l ˙ ) ) = 1 1 − κ r l ( d v h ⃗ d t ⋅ τ r ⃗ + v h ⃗ ⋅ d τ r ⃗ d t ) + 1 1 − κ r l v h ⃗ ⋅ τ r ⃗ 1 − κ r l ( κ r ˙ l + κ r l ˙ ) ) = 1 1 − κ r l ( a ⃗ ⋅ τ r ⃗ + v h ⃗ ⋅ ( κ r s r ˙ n r ⃗ ) ) + 1 1 − κ r l s r ˙ ( d κ r d s r d s r d t l + κ r d l d s r d s r d t ) = a ⃗ ⋅ τ r ⃗ 1 − κ r l + ( κ r s r ˙ ) ( v h ⃗ ⋅ n r ⃗ ) 1 − κ r l + s r ˙ 2 1 − κ r l ( κ r ′ l + κ r l ′ ) (2-15) \begin{aligned} \ddot{s_r} &= \frac{d (\frac{\vec{v_h} \cdot \vec{\tau_r}}{1-\kappa_r l})} {dt} \\ &= \frac{1}{(1-\kappa_r l)^2} (\frac{d(\vec{v_h} \cdot \vec{\tau_r})}{dt}(1-\kappa_r l) - (\vec{v_h} \cdot \vec{\tau_r})(-\dot{\kappa_r}l - \kappa_r \dot{l})) \\ &= \frac{1}{1-\kappa_r l}(\frac{d \vec{v_h}}{dt} \cdot \vec{\tau_r} + \vec{v_h} \cdot \frac{d \vec{\tau_r}}{dt}) + \frac{1}{1-\kappa_r l} \frac{\vec{v_h} \cdot \vec{\tau_r}}{1-\kappa_r l} (\dot{\kappa_r}l + \kappa_r \dot{l})) \\ &= \frac{1}{1-\kappa_r l}(\vec{a} \cdot \vec{\tau_r} + \vec{v_h} \cdot (\kappa_r \dot{s_r} \vec{n_r})) + \frac{1}{1-\kappa_r l} \dot{s_r} (\frac{d \kappa_r}{ds_r} \frac{ds_r}{dt} l + \kappa_r \frac{dl}{ds_r} \frac{ds_r}{dt}) \\ &= \frac{\vec{a} \cdot \vec{\tau_r}}{1-\kappa_r l} + \frac{(\kappa_r \dot{s_r})(\vec{v_h} \cdot \vec{n_r})}{1-\kappa_r l} + \frac{\dot{s_r}^2}{1-\kappa_r l}({\kappa_r}^{\prime}l + \kappa_r l^{\prime}) \end{aligned} \tag{2-15} sr¨=dtd(1κrlvh τr )=(1κrl)21(dtd(vh τr )(1κrl)(vh τr )(κr˙lκrl˙))=1κrl1(dtdvh τr +vh dtdτr )+1κrl11κrlvh τr (κr˙l+κrl˙))=1κrl1(a τr +vh (κrsr˙nr ))+1κrl1sr˙(dsrdκrdtdsrl+κrdsrdldtdsr)=1κrla τr +1κrl(κrsr˙)(vh nr )+1κrlsr˙2(κrl+κrl)(2-15)

由于 l ˙ = v h ⃗ ⋅ n r ⃗ , l ˙ = d l d s d s d t = l ′ s r ˙ \dot{l} = \vec{v_h} \cdot \vec{n_r}, \dot{l} = \frac{dl}{ds} \frac{ds}{dt} = l^{\prime} \dot{s_r} l˙=vh nr ,l˙=dsdldtds=lsr˙,因此:
s r ¨ = a ⃗ ⋅ τ r ⃗ 1 − κ r l + κ r s r ˙ 2 l ′ 1 − κ r l + s r ˙ 2 1 − κ r l ( κ r ′ l + κ r l ′ ) (2-16) \ddot{s_r} = \frac{\vec{a} \cdot \vec{\tau_r}}{1-\kappa_r l} + \frac{\kappa_r \dot{s_r}^2 l^{\prime}}{1-\kappa_r l} + \frac{\dot{s_r}^2}{1-\kappa_r l}({\kappa_r}^{\prime}l + \kappa_r l^{\prime}) \tag{2-16} sr¨=1κrla τr +1κrlκrsr˙2l+1κrlsr˙2(κrl+κrl)(2-16)
根据基础公式⑦: a ⃗ = ∣ v h ˙ ⃗ ∣ τ h ⃗ + ∣ v h ⃗ ∣ 2 κ h n h ⃗ \vec{a} = |\vec{\dot{v_h}}| \vec{\tau_h} + |\vec{v_h}|^2 \kappa_h \vec{n_h} a =vh˙ τh +vh 2κhnh ,由公式 ( 2 − 11 ) (2-11) (211),得: ∣ v h ⃗ ∣ = ( 1 − κ r l ) s r ˙ c o s ( θ h − θ r ) |\vec{v_h}| = \frac{(1-\kappa_r l) \dot{s_r}}{cos(\theta_h - \theta_r)} vh =cos(θhθr)(1κrl)sr˙,因此 a ⃗ ⋅ τ r ⃗ \vec{a} \cdot \vec{\tau_r} a τr 转化为三角函数形式:
a ⃗ ⋅ τ r ⃗ = ∣ v h ˙ ⃗ ∣ c o s < τ h ⃗ , τ r ⃗ > + ∣ v h ⃗ ∣ 2 κ h n h ⃗ ⋅ τ r ⃗ = ∣ v h ˙ ⃗ ∣ c o s < τ h ⃗ , τ r ⃗ > + κ h ( 1 − κ r l ) 2 c o s 2 ( θ h − θ r ) s r ˙ 2 ( − s i n ( θ h − θ r ) ) = ∣ v h ˙ ⃗ ∣ c o s < τ h ⃗ , τ r ⃗ > + κ h s r ˙ 2 1 − κ r l c o s ( θ h − θ r ) ( − ( 1 − κ r l ) t a n ( θ h − θ r ) ) = ∣ a h ⃗ ∣ c o s ( θ h − θ r ) + κ h s r ˙ 2 1 − κ r l c o s ( θ h − θ r ) ( − l ′ ) (2-17) \begin{aligned} \vec{a} \cdot \vec{\tau_r} &= |\vec{\dot{v_h}}| cos<\vec{\tau_h}, \vec{\tau_r}> + |\vec{v_h}|^2 \kappa_h \vec{n_h} \cdot \vec{\tau_r} \\ &= |\vec{\dot{v_h}}| cos<\vec{\tau_h}, \vec{\tau_r}> + \kappa_h \frac{(1-\kappa_r l)^2}{cos^2(\theta_h - \theta_r)} {\dot{s_r}}^2 (-sin(\theta_h - \theta_r)) \\ &= |\vec{\dot{v_h}}| cos<\vec{\tau_h}, \vec{\tau_r}> + \kappa_h {\dot{s_r}}^2 \frac{1-\kappa_r l}{cos(\theta_h - \theta_r)} (-(1-\kappa_r l)tan(\theta_h - \theta_r)) \\ &= |\vec{a_h}| cos(\theta_h - \theta_r) + \kappa_h {\dot{s_r}}^2 \frac{1-\kappa_r l}{cos(\theta_h - \theta_r)} (-l^{\prime}) \end{aligned} \tag{2-17} a τr =vh˙ cos<τh ,τr >+vh 2κhnh τr =vh˙ cos<τh ,τr >+κhcos2(θhθr)(1κrl)2sr˙2(sin(θhθr))=vh˙ cos<τh ,τr >+κhsr˙2cos(θhθr)1κrl((1κrl)tan(θhθr))=ah cos(θhθr)+κhsr˙2cos(θhθr)1κrl(l)(2-17)

因此公式 ( 2 − 16 ) (2-16) (216)的三角函数形式为:
s r ¨ = ∣ a h ⃗ ∣ c o s ( θ h − θ r ) − s r ˙ 2 ( l ′ ( κ h 1 − κ r l c o s ( θ h − θ r ) − κ r ) − ( κ r ′ l + κ r l ′ ) ) 1 − κ r l (2-18) \ddot{s_r} = \frac{|\vec{a_h}| cos(\theta_h - \theta_r) - {\dot{s_r}}^2(l^{\prime} (\kappa_h \frac{1-\kappa_r l}{cos(\theta_h - \theta_r)} - \kappa_r) - ({\kappa_r}^{\prime}l + \kappa_r l^{\prime}))} {1-\kappa_r l} \tag{2-18} sr¨=1κrlah cos(θhθr)sr˙2(l(κhcos(θhθr)1κrlκr)(κrl+κrl))(2-18)

3.3.7 计算 l ¨ \ddot{l} l¨

l ¨ = d l ˙ d t = d v h ⃗ d t ⋅ n r ⃗ + v h ⃗ ⋅ d n r ⃗ d t = a ⃗ ⋅ n r ⃗ + v h ⃗ ⋅ ( − κ r s r ˙ τ r ⃗ ) = a ⃗ ⋅ n r ⃗ − κ r s r ˙ ( v h ⃗ ⋅ τ r ⃗ ) = a ⃗ ⋅ n r ⃗ − κ r ( 1 − κ r l ) ( s r ˙ ) 2 (2-19) \begin{aligned} \ddot{l} &= \frac{d{\dot{l}}}{dt} \\ &= \frac{d \vec{v_h}}{dt} \cdot \vec{n_r} + \vec{v_h} \cdot \frac{d \vec{n_r}}{dt} \\ &= \vec{a} \cdot \vec{n_r} + \vec{v_h} \cdot (- \kappa_r \dot{s_r} \vec{\tau_r}) \\ &= \vec{a} \cdot \vec{n_r} - \kappa_r \dot{s_r} (\vec{v_h} \cdot \vec{\tau_r}) \\ &= \vec{a} \cdot \vec{n_r} - \kappa_r (1-\kappa_r l) (\dot{s_r})^2 \end{aligned} \tag{2-19} l¨=dtdl˙=dtdvh nr +vh dtdnr =a nr +vh (κrsr˙τr )=a nr κrsr˙(vh τr )=a nr κr(1κrl)(sr˙)2(2-19)

a ⃗ ⋅ n r ⃗ \vec{a} \cdot \vec{n_r} a nr 的三角函数形式为:
a ⃗ ⋅ n r ⃗ = ∣ v h ˙ ⃗ ∣ c o s < τ h ⃗ , n r ⃗ > = ∣ v h ˙ ⃗ ∣ c o s ( θ h − ( θ r + π 2 ) ) = ∣ a ⃗ ∣ s i n ( θ h − θ r ) (2-20) \vec{a} \cdot \vec{n_r} = |\vec{\dot{v_h}}| cos<\vec{\tau_h}, \vec{n_r}> = |\vec{\dot{v_h}}| cos(\theta_h - (\theta_r + \frac{\pi}{2})) = |\vec{a}| sin(\theta_h - \theta_r) \tag{2-20} a nr =vh˙ cos<τh ,nr >=vh˙ cos(θh(θr+2π))=a sin(θhθr)(2-20)

因此, l ¨ \ddot{l} l¨的三角函数形式为:
l ¨ = ∣ a ⃗ ∣ s i n ( θ h − θ r ) − κ r ( 1 − κ r l ) ( s r ˙ ) 2 (2-21) \ddot{l} = |\vec{a}| sin(\theta_h - \theta_r) - \kappa_r (1-\kappa_r l) (\dot{s_r})^2 \tag{2-21} l¨=a sin(θhθr)κr(1κrl)(sr˙)2(2-21)

3.3.8 计算 l ′ ′ l^{\prime \prime} l

l ¨ = d l ˙ d t = d l ′ s r ˙ ˙ d t = d l ′ d s r d s r d t s r ˙ + l ′ s r ¨ = l ′ ′ s r ˙ 2 + l ′ s r ¨ (2-22) \begin{aligned} \ddot{l} &= \frac{d{\dot{l}}}{dt} \\ &= \frac{d{\dot{l^{\prime} \dot{s_r}}}}{dt} \\ &= \frac{d {l^{\prime}}}{ds_r} \frac{ds_r}{dt} \dot{s_r} + l^{\prime} \ddot{s_r} \\ &= l^{\prime \prime} {\dot{s_r}}^2 + l^{\prime} \ddot{s_r} \end{aligned} \tag{2-22} l¨=dtdl˙=dtdlsr˙˙=dsrdldtdsrsr˙+lsr¨=lsr˙2+lsr¨(2-22)

因此,可以得到:
l ′ ′ = l ¨ − l ′ s r ¨ s r ˙ 2 (2-23) l^{\prime \prime} = \frac{\ddot{l} - l^{\prime} \ddot{s_r}}{{\dot{s_r}}^2} \tag{2-23} l=sr˙2l¨lsr¨(2-23)

写成三角函数形式:
l ′ ′ = − ( κ r ′ l + κ r l ′ ) t a n ( θ h − θ r ) + 1 − κ r l c o s 2 ( θ h − θ r ) ( 1 − κ r l c o s ( θ h − θ r ) κ h − κ r ) (2-24) l^{\prime \prime} = -({\kappa_r}^{\prime}l + \kappa_r l^{\prime}) tan(\theta_h - \theta_r) + \frac{1-\kappa_r l}{cos^2(\theta_h - \theta_r)} (\frac{1-\kappa_r l}{cos(\theta_h - \theta_r)} \kappa_h - \kappa_r) \tag{2-24} l=(κrl+κrl)tan(θhθr)+cos2(θhθr)1κrl(cos(θhθr)1κrlκhκr)(2-24)

Reference

[1] 自动驾驶决策规划算法第一章第三节(上) 直角坐标与自然坐标转换

[2] Frenet坐标系与Cartesian坐标系互转

  • 3
    点赞
  • 7
    收藏
    觉得还不错? 一键收藏
  • 2
    评论
评论 2
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值