来自论文"Predictive Path-Following Control for Fixed-Wing UAVs Using the qLMPC Framework in the Presence of Wind Disturbances"
控制架构
采用的是 ULTRA-Extra无人机,相关参数如下:
这里用于guidance law的无人机运动学模型为:
{ x ˙ p = V a cos γ cos χ + V w cos γ w cos χ w y ˙ p = V a cos γ sin χ + V w cos γ w sin χ w z ˙ p = V a sin γ + V w sin γ w χ ˙ = g tan ϕ / V a γ ˙ = g ( n z cos ϕ − cos γ ) / V a \begin{cases} \dot{x}_p = V_a\cos\gamma\cos\chi + V_w\cos\gamma_w\cos\chi_w \\ \dot{y}_p = V_a\cos\gamma\sin\chi + V_w\cos\gamma_w\sin\chi_w \\ \dot{z}_p = V_a\sin\gamma + V_w\sin\gamma_w \\ \dot{\chi} = g\tan\phi/V_a \\ \dot{\gamma} = g(n_z\cos\phi-\cos\gamma)/V_a \end{cases} ⎩ ⎨ ⎧x˙p=Vacosγcosχ+Vwcosγwcosχwy˙p=Vacosγsinχ+Vwcosγwsinχwz˙p=Vasinγ+Vwsinγwχ˙=gtanϕ/Vaγ˙=g(nzcosϕ−cosγ)/Va
其中状态量为 ( x p , y p , z p , γ , χ ) (x_p,y_p,z_p,\gamma,\chi) (xp,yp,zp,γ,χ),控制量为 ( V a , n z , ϕ ) (V_a,n_z,\phi) (Va,nz,ϕ)。在自动驾驶仪(Autopilot)中,采用 Successive-Loop-Closure (SLC)实现参考量 ( V a m , n z m , ϕ m ) (V_{a_m},n_{z_m},\phi_m) (Vam,nzm,ϕm)的信号跟踪:
自动驾驶仪中依然采用横纵向通道的SLC实现控制,相应的控制逻辑如下:
Path Following 最优控制器
对运动学模型进行二阶求导可以得到:
( x ˙ p y ˙ p z ˙ p χ ˙ γ ˙ x ¨ p y ¨ p z ¨ p χ ¨ γ ¨ ) = ( O 5 × 5 I 5 − V a cos γ sin χ − V a sin γ cos χ V a cos γ cos χ − V a sin γ sin χ O 5 × 5 O 5 × 3 0 V a cos γ 0 0 0 g sin γ V a ) ( x p y p z p χ γ x ˙ p y ˙ p z ˙ p χ ˙ γ ˙ ) + ( O 5 × 3 cos γ cos χ 0 0 cos γ sin χ 0 0 sin γ 0 0 − g tan ϕ V a 2 g V a cos 2 ϕ 0 g ( cos γ − n z cos ϕ ) V a 2 − g n z sin ϕ V a g cos ϕ V a ) ( V ˙ a ϕ ˙ n ˙ z ) \left( \begin{matrix} {{{\dot{x}}}_{p}} \\ {{{\dot{y}}}_{p}} \\ {{{\dot{z}}}_{p}} \\ {\dot{\chi }} \\ {\dot{\gamma }} \\ {{{\ddot{x}}}_{p}} \\ {{{\ddot{y}}}_{p}} \\ {{{\ddot{z}}}_{p}} \\ {\ddot{\chi }} \\ {\ddot{\gamma }} \\ \end{matrix} \right)=\left( \begin{matrix} {{O}_{5\times 5}} & {} & {{I}_{5}} & {} \\ {} & {} & -{{V}_{a}}\cos \gamma \sin \chi & -{{V}_{a}}\sin \gamma \cos \chi \\ {} & {} & {{V}_{a}}\cos \gamma \cos \chi & -{{V}_{a}}\sin \gamma \sin \chi \\ {{O}_{5\times 5}} & {{O}_{5\times 3}} & 0 & {{V}_{a}}\cos \gamma \\ {} & {} & 0 & 0 \\ {} & {} & 0 & \frac{g\sin \gamma }{V_{a}^{{}}} \\ \end{matrix} \right)\left( \begin{matrix} {{x}_{p}} \\ {{y}_{p}} \\ {{z}_{p}} \\ \chi \\ \gamma \\ {{{\dot{x}}}_{p}} \\ {{{\dot{y}}}_{p}} \\ {{{\dot{z}}}_{p}} \\ {\dot{\chi }} \\ {\dot{\gamma }} \\ \end{matrix} \right)+\left( \begin{matrix} {} & {{O}_{5\times 3}} & {} \\ \cos \gamma \cos \chi & 0 & 0 \\ \cos \gamma \sin \chi & 0 & 0 \\ \sin \gamma & 0 & 0 \\ -\frac{g\tan \phi }{V_{a}^{2}} & \frac{g}{{{V}_{a}}{{\cos }^{2}}\phi } & 0 \\ \frac{g(\cos \gamma -{{n}_{z}}\cos \phi )}{V_{a}^{2}} & -\frac{g{{n}_{z}}\sin \phi }{V_{a}^{{}}} & \frac{g\cos \phi }{V_{a}^{{}}} \\ \end{matrix} \right)\left( \begin{align} & {{{\dot{V}}}_{a}} \\ & {\dot{\phi }} \\ & {{{\dot{n}}}_{z}} \\ \end{align} \right) x˙py˙pz˙pχ˙γ˙x¨py¨pz¨pχ¨γ¨ = O5×5O5×5O5×3I5−VacosγsinχVacosγcosχ000−Vasinγcosχ−VasinγsinχVacosγ0Vagsinγ xpypzpχγx˙py˙pz˙pχ˙γ˙ + cosγcosχcosγsinχsinγ−Va2gtanϕVa2g(cosγ−nzcosϕ)O5×3000Vacos2ϕg−Vagnzsinϕ0000Vagcosϕ V˙aϕ˙n˙z
这里设 ρ = ( γ , χ , V a , ϕ , n z ) T \rho=(\gamma,\chi,V_a,\phi,n_z)^T ρ=(γ,χ,Va,ϕ,nz)T, x = ( x p , y p , z p , χ , γ , x ˙ p , y ˙ p , z ˙ p , χ ˙ , γ ˙ ) T x=(x_p,y_p,z_p,\chi,\gamma,\dot{x}_p,\dot{y}_p,\dot{z}_p,\dot{\chi},\dot{\gamma})^T x=(xp,yp,zp,χ,γ,x˙p,y˙p,z˙p,χ˙,γ˙)T, u = ( V ˙ a , ϕ ˙ , n ˙ z ) T u=(\dot{V}_a,\dot{\phi},\dot{n}_z)^T u=(V˙a,ϕ˙,n˙z)T,得到:
x ˙ = A v ( ρ ) x + B v ( ρ ) u \dot{x}=A_v(\rho)x+B_v(\rho)u x˙=Av(ρ)x+Bv(ρ)u
假设要跟踪的量为 r = ( x r , y r , z r ) T r=(x_r,y_r,z_r)^T r=(xr,yr,zr)T,构造跟踪向量 e = ( x r − x p , y r − y p , z r − z p ) T = r − ( x p , y p , z p ) T e=(x_r-x_p,y_r-y_p,z_r-z_p)^T=r-(x_p,y_p,z_p)^T e=(xr−xp,yr−yp,zr−zp)T=r−(xp,yp,zp)T,有: e ˙ = r ˙ − ( O 3 × 5 , I 3 , O 3 × 2 ) x \dot{e}=\dot{r}-(O_{3\times 5},I_3,O_{3\times 2})x e˙=r˙−(O3×5,I3,O3×2)x,得到:
( x ˙ e ˙ ) = ( A v ( ρ ) O 10 × 3 − ( O 3 × 5 ∣ I 3 ∣ O 3 × 2 ) O 3 ) ( x e ) + ( B v ( ρ ) O 3 × 3 ) u + ( O 10 × 1 r ˙ ) \begin{pmatrix} \dot{x} \\ \dot{e} \end{pmatrix} = \begin{pmatrix} A_v(\rho) & O_{10\times3} \\ -(O_{3\times 5}|I_3|O_{3\times 2}) & O_3 \end{pmatrix}\begin{pmatrix} x \\e \end{pmatrix} + \begin{pmatrix} B_v(\rho) \\O_{3\times 3} \end{pmatrix} u + \begin{pmatrix} O_{10\times 1} \\\dot{r} \end{pmatrix} (x˙e˙)=(Av(ρ)−(O3×5∣I3∣O3×2)O10×3O3)(xe)+(Bv(ρ)O3×3)u+(O10×1r˙)
利用4阶Runge-Kutta法可以将上式可以离散化为一个LPV状态空间方程(linear parameter varying state-space representation):
x e , k + 1 = A e ( ρ k ) x e , k + B e ( ρ k ) u e , k + c r , k x_{e,k+1} = A_e(\rho_k)x_{e,k}+B_e(\rho_k)u_{e,k}+c_{r,k} xe,k+1=Ae(ρk)xe,k+Be(ρk)ue,k+cr,k
令 P k = ( ρ k T , ρ k + 1 T , . . . ρ k + N − 1 T ) T P_k=(\rho_{k}^T,\rho_{k+1}^T,...\rho_{k+N-1}^T)^T Pk=(ρkT,ρk+1T,...ρk+N−1T)T, X k = ( x e , k + 1 T , x e , k + 2 T , . . . , x e , k + N T ) T X_k=(x_{e,k+1}^T,x_{e,k+2}^T,...,x_{e,k+N}^T)^T Xk=(xe,k+1T,xe,k+2T,...,xe,k+NT)T, U k = ( u e , k + 1 T , u e , k + 2 T , . . . , u e , k + N T ) T U_k=(u_{e,k+1}^T,u_{e,k+2}^T,...,u_{e,k+N}^T)^T Uk=(ue,k+1T,ue,k+2T,...,ue,k+NT)T, c k = ( c r , k + 1 T , c r , k + 2 T , . . . c r , k + N T ) T c_k = (c_{r,k+1}^T,c_{r,k+2}^T,...c_{r,k+N}^T)^T ck=(cr,k+1T,cr,k+2T,...cr,k+NT)T得到:
X k + 1 = H ( P k ) X k + S ( P k ) U k + c k = L k + S ( P k ) U k X_{k+1}=H(P_k)X_k + S(P_k)U_k+c_k \\ =L_k + S(P_k)U_k Xk+1=H(Pk)Xk+S(Pk)Uk+ck=Lk+S(Pk)Uk
其中: x e , k = ( x k T , e k T ) T x_{e,k}=(x^T_k,e^T_k)^T xe,k=(xkT,ekT)T, H ( P k ) = d i a g ( [ A e ( ρ k ) , A e ( ρ k + 1 ) . . . A e ( ρ k + N − 1 ) ] ) H(P_k) =diag([A_e(\rho_{k}),A_e(\rho_{k+1})...A_e(\rho_{k+N-1})]) H(Pk)=diag([Ae(ρk),Ae(ρk+1)...Ae(ρk+N−1)]), S ( P k ) = d i a g ( [ B e ( ρ k ) , B e ( ρ k + 1 ) . . . B e ( ρ k + N − 1 ) ] ) S(P_k)=diag([B_e(\rho_{k}),B_e(\rho_{k+1})...B_e(\rho_{k+N-1})]) S(Pk)=diag([Be(ρk),Be(ρk+1)...Be(ρk+N−1)])。采用MPC控制器进行设计时, k + 1 k+1 k+1时刻需要优化的目标函数:
J k + 1 = ∑ i = 1 N ( x k + i + 1 T Q x k + i + 1 + u k + i T R u k + i + e k + i + 1 T T e k + i + 1 ) = X k + 1 T H X X k + 1 + U k T H U U k = [ L k + S ( P k ) U k ] T H X [ L k + S ( P k ) U k ] + U k T H U U k = U k T ( S ( P k ) T H X S ( P k ) + H U ) U k + 2 L k T H X S ( P k ) U k + L k T H X L k J_{k+1}=\sum_{i=1}^N(x_{k+i+1}^TQx_{k+i+1} + u_{k+i}^TRu_{k+i} + e_{k+i+1}^TTe_{k+i+1}) \\ =X_{k+1}^TH_XX_{k+1} + U_{k}^TH_UU_{k}\\=[L_k + S(P_k)U_k]^TH_X[L_k + S(P_k)U_k] + U_{k}^TH_UU_{k}\\ =U_k^T(S(P_k)^TH_XS(P_k)+H_U)U_k + 2L_k^TH_XS(P_k)U_k + L_k^TH_XL_k Jk+1=i=1∑N(xk+i+1TQxk+i+1+uk+iTRuk+i+ek+i+1TTek+i+1)=Xk+1THXXk+1+UkTHUUk=[Lk+S(Pk)Uk]THX[Lk+S(Pk)Uk]+UkTHUUk=UkT(S(Pk)THXS(Pk)+HU)Uk+2LkTHXS(Pk)Uk+LkTHXLk
其中: Q = Q T > 0 , P = P T > 0 , R = R T > 0 Q=Q^T>0,P=P^T>0,R=R^T>0 Q=QT>0,P=PT>0,R=RT>0; H X = d i a g ( [ Q , Q , . . . , Q ] ) H_X=diag([Q,Q,...,Q]) HX=diag([Q,Q,...,Q]), H U = d i a g ( [ R , R , . . . , R ] ) H_U=diag([R,R,...,R]) HU=diag([R,R,...,R])。
而对于控制量 U k U_k Uk和状态量 X k + 1 X_{k+1} Xk+1有限幅,即: U min ≤ U k ≤ U max U_{\min}\leq U_k\leq U_{\max} Umin≤Uk≤Umax, X min ≤ X k + 1 ≤ X max X_{\min} \leq X_{k+1} \leq X_{\max} Xmin≤Xk+1≤Xmax,得到约束:
( I − I S ( P k ) − S ( P k ) ) U k ≤ ( U max − U min X k + 1 − L k − X k + 1 + L k ) \begin{pmatrix} I \\ -I\\ S(P_k)\\ -S(P_k) \end{pmatrix}U_k \leq \begin{pmatrix} U_{\max} \\ -U_{\min} \\ X_{k+1} - L_k \\ -X_{k+1} + L_k \end{pmatrix} I−IS(Pk)−S(Pk) Uk≤ Umax−UminXk+1−Lk−Xk+1+Lk
上面的假设是基于全状态反馈的,也是就是说对于 k + 1 k+1 k+1时刻的在线优化能获取 k k k时刻所有的状态信息和偏差信息。
若观测量为 Y k = C k X k Y_k = C_kX_k Yk=CkXk, Y min ≤ Y k ≤ Y max Y_{\min}\leq Y_k\leq Y_{\max} Ymin≤Yk≤Ymax,则上面的约束将被修正为:
( I − I C k S ( P k ) − C k S ( P k ) ) U k ≤ ( U max − U min Y max − C k L k − Y min + C k L k ) \begin{pmatrix} I \\ -I\\ C_kS(P_k)\\ -C_kS(P_k) \end{pmatrix}U_k \leq \begin{pmatrix} U_{\max} \\ -U_{\min} \\ Y_{\max} - C_kL_k \\ -Y_{\min} + C_kL_k \end{pmatrix} I−ICkS(Pk)−CkS(Pk) Uk≤ Umax−UminYmax−CkLk−Ymin+CkLk
无论如何,上述问题都可以被转化成QP问题,利用Matlab工具箱中的quadprog
函数进行求解,或者说是在线优化为以下问题:
min U k 1 2 U k T F k U k + G k U k A k U k ≤ b k \min_{U_k}\frac{1}{2}U_k^TF_kU_k +G_kU_k \\ A_kU_k \leq b_k Ukmin21UkTFkUk+GkUkAkUk≤bk
附带相应的伪代码如下图所示:
参考文献
@inproceedings{bib:Samir,title={Predictive Path Following Control for Fixed Wing UAVs Using the qLMPC Framework in the Presence of Wind Disturbances},author={Rezk, Ahmed S and Calder{\'o}n, Horacio M and Werner, Herbert and Herrmann, Benjamin and Thielecke, Frank},booktitle={AIAA SCITECH 2024 Forum},pages={1594},year={2024}
}