added: parse Hc argument for move blocking

Author: franckgaga

src/controller/construct.jl

@@ -453,6 +453,12 @@ end
 "Return `0` when model is not a [`LinModel`](@ref)."
 estimate_delays(::SimModel) = 0
 
+
+"Get move blocking vector `nb` and actual `Hc` value from provided `Hc_arg` argument" 
+move_blocking(Hc_arg::AbstractVector{Int}) = (Hc_arg, length(nb))
+move_blocking(Hc_arg::Int) = (fill(1, length(Hc_arg)), Hc_arg)
+
+
 """
     validate_args(mpc::PredictiveController, ry, d, D̂, R̂y, R̂u)
 

src/controller/explicitmpc.jl

@@ -39,7 +39,7 @@ struct ExplicitMPC{
     Dop::Vector{NT}
     buffer::PredictiveControllerBuffer{NT}
     function ExplicitMPC{NT}(
-        estim::SE, Hp, Hc, weights::CW
+        estim::SE, Hp, Hc, nb, weights::CW
     ) where {NT<:Real, SE<:StateEstimator, CW<:ControllerWeights}
         model = estim.model
         nu, ny, nd, nx̂ = model.nu, model.ny, model.nd, estim.nx̂
@@ -48,7 +48,6 @@ struct ExplicitMPC{
         # dummy vals (updated just before optimization):
         R̂y, R̂u, Tu_lastu0 = zeros(NT, ny*Hp), zeros(NT, nu*Hp), zeros(NT, nu*Hp)
         transcription = SingleShooting() # explicit MPC only supports SingleShooting
-        nb = ones(Hc)
         PΔu = init_ZtoΔU(estim, transcription, Hp, Hc)
         Pu, Tu = init_ZtoU(estim, transcription, Hp, Hc)
         E, G, J, K, V, B = init_predmat(model, estim, transcription, Hp, Hc)
@@ -128,7 +127,7 @@ ExplicitMPC controller with a sample time Ts = 4.0 s, SteadyKalmanFilter estimat
 function ExplicitMPC(
     model::LinModel; 
     Hp::Int = default_Hp(model),
-    Hc::Int = DEFAULT_HC,
+    Hc::IntVectorOrInt = DEFAULT_HC,
     Mwt = fill(DEFAULT_MWT, model.ny),
     Nwt = fill(DEFAULT_NWT, model.nu),
     Lwt = fill(DEFAULT_LWT, model.nu),
@@ -151,7 +150,7 @@ Use custom state estimator `estim` to construct `ExplicitMPC`.
 function ExplicitMPC(
     estim::SE;
     Hp::Int = default_Hp(estim.model),
-    Hc::Int = DEFAULT_HC,
+    Hc::IntVectorOrInt = DEFAULT_HC,
     Mwt  = fill(DEFAULT_MWT, estim.model.ny),
     Nwt  = fill(DEFAULT_NWT, estim.model.nu),
     Lwt  = fill(DEFAULT_LWT, estim.model.nu),
@@ -165,8 +164,9 @@ function ExplicitMPC(
         @warn("prediction horizon Hp ($Hp) ≤ estimated number of delays in model "*
               "($nk), the closed-loop system may be unstable or zero-gain (unresponsive)")
     end
+    nb, Hc = move_blocking(Hc)
     weights = ControllerWeights{NT}(estim.model, Hp, Hc, M_Hp, N_Hc, L_Hp)
-    return ExplicitMPC{NT}(estim, Hp, Hc, weights)
+    return ExplicitMPC{NT}(estim, Hp, Hc, nb, weights)
 end
 
 setconstraint!(::ExplicitMPC; kwargs...) = error("ExplicitMPC does not support constraints.")

src/controller/linmpc.jl

@@ -47,7 +47,7 @@ struct LinMPC{
     Dop::Vector{NT}
     buffer::PredictiveControllerBuffer{NT}
     function LinMPC{NT}(
-        estim::SE, Hp, Hc, weights::CW, 
+        estim::SE, Hp, Hc, nb, weights::CW, 
         transcription::TM, optim::JM
     ) where {
             NT<:Real, 
@@ -61,7 +61,6 @@ struct LinMPC{
         ŷ = copy(model.yop) # dummy vals (updated just before optimization)
         # dummy vals (updated just before optimization):
         R̂y, R̂u, Tu_lastu0 = zeros(NT, ny*Hp), zeros(NT, nu*Hp), zeros(NT, nu*Hp)
-        nb = ones(Hc)
         PΔu = init_ZtoΔU(estim, transcription, Hp, Hc)
         Pu, Tu = init_ZtoU(estim, transcription, Hp, Hc)
         E, G, J, K, V, B, ex̂, gx̂, jx̂, kx̂, vx̂, bx̂ = init_predmat(
@@ -204,7 +203,7 @@ LinMPC controller with a sample time Ts = 4.0 s, OSQP optimizer, SteadyKalmanFil
 function LinMPC(
     model::LinModel;
     Hp::Int = default_Hp(model),
-    Hc::Int = DEFAULT_HC,
+    Hc::IntVectorOrInt = DEFAULT_HC,
     Mwt  = fill(DEFAULT_MWT, model.ny),
     Nwt  = fill(DEFAULT_NWT, model.nu),
     Lwt  = fill(DEFAULT_LWT, model.nu),
@@ -247,7 +246,7 @@ LinMPC controller with a sample time Ts = 4.0 s, OSQP optimizer, KalmanFilter es
 function LinMPC(
     estim::SE;
     Hp::Int = default_Hp(estim.model),
-    Hc::Int = DEFAULT_HC,
+    Hc::IntVectorOrInt = DEFAULT_HC,
     Mwt  = fill(DEFAULT_MWT, estim.model.ny),
     Nwt  = fill(DEFAULT_NWT, estim.model.nu),
     Lwt  = fill(DEFAULT_LWT, estim.model.nu),
@@ -264,8 +263,9 @@ function LinMPC(
         @warn("prediction horizon Hp ($Hp) ≤ estimated number of delays in model "*
               "($nk), the closed-loop system may be unstable or zero-gain (unresponsive)")
     end
+    nb, Hc = move_blocking(Hc)
     weights = ControllerWeights{NT}(estim.model, Hp, Hc, M_Hp, N_Hc, L_Hp, Cwt)
-    return LinMPC{NT}(estim, Hp, Hc, weights, transcription, optim)
+    return LinMPC{NT}(estim, Hp, Hc, nb, weights, transcription, optim)
 end
 
 """

src/controller/nonlinmpc.jl

@@ -63,7 +63,7 @@ struct NonLinMPC{
     Dop::Vector{NT}
     buffer::PredictiveControllerBuffer{NT}
     function NonLinMPC{NT}(
-        estim::SE, Hp, Hc, weights::CW,
+        estim::SE, Hp, Hc, nb, weights::CW,
         JE::JEfunc, gc!::GCfunc, nc, p::PT, 
         transcription::TM, optim::JM, 
         gradient::GB, jacobian::JB
@@ -84,7 +84,6 @@ struct NonLinMPC{
         ŷ = copy(model.yop) # dummy vals (updated just before optimization)
         # dummy vals (updated just before optimization):
         R̂y, R̂u, Tu_lastu0 = zeros(NT, ny*Hp), zeros(NT, nu*Hp), zeros(NT, nu*Hp)
-        nb = ones(Hc)
         PΔu = init_ZtoΔU(estim, transcription, Hp, Hc)
         Pu, Tu = init_ZtoU(estim, transcription, Hp, Hc)
         E, G, J, K, V, B, ex̂, gx̂, jx̂, kx̂, vx̂, bx̂ = init_predmat(
@@ -280,7 +279,7 @@ NonLinMPC controller with a sample time Ts = 10.0 s, Ipopt optimizer, UnscentedK
 function NonLinMPC(
     model::SimModel;
     Hp::Int = default_Hp(model),
-    Hc::Int = DEFAULT_HC,
+    Hc::IntVectorOrInt = DEFAULT_HC,
     Mwt  = fill(DEFAULT_MWT, model.ny),
     Nwt  = fill(DEFAULT_NWT, model.nu),
     Lwt  = fill(DEFAULT_LWT, model.nu),
@@ -337,7 +336,7 @@ NonLinMPC controller with a sample time Ts = 10.0 s, Ipopt optimizer, UnscentedK
 function NonLinMPC(
     estim::SE;
     Hp::Int = default_Hp(estim.model),
-    Hc::Int = DEFAULT_HC,
+    Hc::IntVectorOrInt = DEFAULT_HC,
     Mwt  = fill(DEFAULT_MWT, estim.model.ny),
     Nwt  = fill(DEFAULT_NWT, estim.model.nu),
     Lwt  = fill(DEFAULT_LWT, estim.model.nu),
@@ -364,11 +363,12 @@ function NonLinMPC(
         @warn("prediction horizon Hp ($Hp) ≤ estimated number of delays in model "*
               "($nk), the closed-loop system may be unstable or zero-gain (unresponsive)")
     end
+    nb, Hc = move_blocking(Hc)
     validate_JE(NT, JE)
     gc! = get_mutating_gc(NT, gc)
     weights = ControllerWeights{NT}(estim.model, Hp, Hc, M_Hp, N_Hc, L_Hp, Cwt, Ewt)
     return NonLinMPC{NT}(
-        estim, Hp, Hc, weights, JE, gc!, nc, p, transcription, optim, gradient, jacobian
+        estim, Hp, Hc, nb, weights, JE, gc!, nc, p, transcription, optim, gradient, jacobian
     )
 end
 

src/controller/transcription.jl

@@ -10,7 +10,8 @@ abstract type TranscriptionMethod end
 
 Construct a direct single shooting [`TranscriptionMethod`](@ref).
 
-The decision variable in the optimization problem is (excluding the slack ``ϵ``):
+The decision variable in the optimization problem is (excluding the slack ``ϵ`` and without
+any custom move blocking):
 ```math
 \mathbf{Z} = \mathbf{ΔU} =          \begin{bmatrix} 
     \mathbf{Δu}(k+0)                \\ 
@@ -66,7 +67,7 @@ end
 Init decision variables to input increments over ``H_c`` conversion matrix `PΔu`.
 
 Introducing the move blocking vector that specifies the length of each step in the 
-``\mathbf{ΔU}`` vector to customize the pattern (strictly positive integers):
+``\mathbf{ΔU}`` vector, to customize the pattern (strictly positive integers):
 
 ```math
 \mathbf{n_b} = \begin{bmatrix} n_1 & n_2 & \cdots & n_{H_c} \end{bmatrix}'