debug: ExplicitMPC correct error message for setconstraint!

improve coverage
ignore `lcov.info` file

Author: franckgaga

.gitignore

@@ -5,3 +5,4 @@ LocalPreferences.toml
 .vscode/
 *.cov
 docs/Manifest.toml
+lcov.info

Project.toml

@@ -1,7 +1,7 @@
 name = "ModelPredictiveControl"
 uuid = "61f9bdb8-6ae4-484a-811f-bbf86720c31c"
 authors = ["Francis Gagnon"]
-version = "0.24.1"
+version = "1.0.0"
 
 [deps]
 ControlSystemsBase = "aaaaaaaa-a6ca-5380-bf3e-84a91bcd477e"

README.md

@@ -24,7 +24,7 @@ using Pkg; Pkg.add("ModelPredictiveControl")
 To construct model predictive controllers (MPCs), we must first specify a plant model that
 is typically extracted from input-output data using [system identification](https://github.com/baggepinnen/ControlSystemIdentification.jl).
 The model here is linear with one input, two outputs and a large time delay in the first
-channel:
+channel (a transfer matrix, with $s$ as the Laplace variable):
 
 ```math
 \mathbf{G}(s) = \frac{\mathbf{y}(s)}{\mathbf{u}(s)} = 

src/controller/explicitmpc.jl

@@ -167,7 +167,7 @@ function ExplicitMPC(
     return ExplicitMPC{NT, SE}(estim, Hp, Hc, M_Hp, N_Hc, L_Hp)
 end
 
-setconstraint!(::ExplicitMPC,kwargs...) = error("ExplicitMPC does not support constraints.")
+setconstraint!(::ExplicitMPC; kwargs...) = error("ExplicitMPC does not support constraints.")
 
 function Base.show(io::IO, mpc::ExplicitMPC)
     Hp, Hc = mpc.Hp, mpc.Hc

src/model/solver.jl

@@ -16,6 +16,9 @@ struct RungeKutta <: DiffSolver
         if order ≠ 4
             throw(ArgumentError("only 4th order Runge-Kutta is supported."))
         end
+        if order < 1
+            throw(ArgumentError("order must be greater than 0"))
+        end
         if supersample < 1
             throw(ArgumentError("supersample must be greater than 0"))
         end
@@ -31,15 +34,7 @@ Create a Runge-Kutta solver with optional super-sampling.
 Only the 4th order Runge-Kutta is supported for now. The keyword argument `supersample`
 provides the number of internal steps (default to 1 step).
 """
-function RungeKutta(order::Int=4; supersample::Int=1)
-    if order < 1
-        throw(ArgumentError("order must be greater than 0"))
-    end
-    if supersample < 1
-        throw(ArgumentError("supersample must be greater than 0"))
-    end
-    return RungeKutta(order, supersample)
-end
+RungeKutta(order::Int=4; supersample::Int=1) = RungeKutta(order, supersample)
 
 "Get the `f!` and `h!` functions for Runge-Kutta solver."
 function get_solver_functions(NT::DataType, solver::RungeKutta, fc!, hc!, Ts, _ , nx, _ , _ )

src/sim_model.jl

@@ -141,19 +141,19 @@ LinModel with a sample time Ts = 2.0 s and:
 """
 function setname!(model::SimModel; u=nothing, y=nothing, d=nothing, x=nothing)
     if !isnothing(u)
-        size(u) == (model.nu,) || error("u size must be $((model.nu,))")
+        size(u) == (model.nu,) || throw(DimensionMismatch("u size must be $((model.nu,))"))
         model.uname .= u
     end
     if !isnothing(y)
-        size(y) == (model.ny,) || error("y size must be $((model.ny,))")
+        size(y) == (model.ny,) || throw(DimensionMismatch("y size must be $((model.ny,))"))
         model.yname .= y
     end
     if !isnothing(d)
-        size(d) == (model.nd,) || error("d size must be $((model.nd,))")
+        size(d) == (model.nd,) || throw(DimensionMismatch("d size must be $((model.nd,))"))
         model.dname .= d
     end
     if !isnothing(x)
-        size(x) == (model.nx,) || error("x size must be $((model.nx,))")
+        size(x) == (model.nx,) || throw(DimensionMismatch("x size must be $((model.nx,))"))
         model.xname .= x
     end
     return model

test/test_plot_sim.jl

@@ -110,6 +110,18 @@ end
     @test all(p5[end-4][2][:y] .≈ +101)
     @test all(p5[end-3][2][:y] .≈ +102)
     @test all(p5[end-2][2][:y] .≈ +103)
+    p6 = plot(res, plotxwithx̂=true, plotx̂min=true, plotx̂max=false)
+    @test p6[1][1][:x] ≈ res.T_data
+    @test all(p6[end-5][3][:y] .≈ -100)
+    @test all(p6[end-4][3][:y] .≈ -101)
+    @test all(p6[end-3][3][:y] .≈ -102)
+    @test all(p6[end-2][3][:y] .≈ -103)
+    p7 = plot(res, plotxwithx̂=true, plotx̂min=false, plotx̂max=true)
+    @test p7[1][1][:x] ≈ res.T_data
+    @test all(p7[end-5][3][:y] .≈ +100)
+    @test all(p7[end-4][3][:y] .≈ +101)
+    @test all(p7[end-3][3][:y] .≈ +102)
+    @test all(p7[end-2][3][:y] .≈ +103)
 end
 
 @testset "PredictiveController quick simulation" begin
@@ -147,7 +159,10 @@ end
 end
 
 @testset "PredictiveController Plots" begin
-    mpc = LinMPC(LinModel(sys, Ts, i_d=[3]), Lwt=[0.01, 0.01])
+    estim = MovingHorizonEstimator(LinModel(sys, Ts, i_d=[3]), He=5)
+    estim = setconstraint!(estim, x̂min=[-100,-101,-102,-103,-104,-105])
+    estim = setconstraint!(estim, x̂max=[+100,+101,+102,+103,+104,+105])
+    mpc = LinMPC(estim, Lwt=[0.01, 0.01])
     mpc = setconstraint!(mpc, umin=[-50, -51], umax=[52, 53], ymin=[-54,-55], ymax=[56,57])
     res = sim!(mpc, 15)
     p1 = plot(res, plotŷ=true)
@@ -198,4 +213,36 @@ end
     @test p8[1][1][:x] ≈ res.T_data
     @test all(p8[end-4][3][:y] .≈ 56.0)
     @test all(p8[end-3][3][:y] .≈ 57.0)
+    p9 = plot(res, plotx̂=true, plotx̂min=true, plotx̂max=false)
+    @test p9[1][1][:x] ≈ res.T_data
+    @test all(p9[end-5][2][:y] .≈ -100.0)
+    @test all(p9[end-4][2][:y] .≈ -101.0)
+    @test all(p9[end-3][2][:y] .≈ -102.0)
+    @test all(p9[end-2][2][:y] .≈ -103.0)
+    @test all(p9[end-1][2][:y] .≈ -104.0)
+    @test all(p9[end-0][2][:y] .≈ -105.0)
+    p10 = plot(res, plotx̂=true, plotx̂min=false, plotx̂max=true)
+    @test p10[1][1][:x] ≈ res.T_data
+    @test all(p10[end-5][2][:y] .≈ +100.0)
+    @test all(p10[end-4][2][:y] .≈ +101.0)
+    @test all(p10[end-3][2][:y] .≈ +102.0)
+    @test all(p10[end-2][2][:y] .≈ +103.0)
+    @test all(p10[end-1][2][:y] .≈ +104.0)
+    @test all(p10[end-0][2][:y] .≈ +105.0)
+    p11 = plot(res, plotxwithx̂=true, plotx̂min=true, plotx̂max=false)
+    @test p11[1][1][:x] ≈ res.T_data
+    @test all(p11[end-5][3][:y] .≈ -100.0)
+    @test all(p11[end-4][3][:y] .≈ -101.0)
+    @test all(p11[end-3][3][:y] .≈ -102.0)
+    @test all(p11[end-2][3][:y] .≈ -103.0)
+    @test all(p11[end-1][2][:y] .≈ -104.0)
+    @test all(p11[end-0][2][:y] .≈ -105.0)
+    p12 = plot(res, plotxwithx̂=true, plotx̂min=false, plotx̂max=true)
+    @test p12[1][1][:x] ≈ res.T_data
+    @test all(p12[end-5][3][:y] .≈ +100.0)
+    @test all(p12[end-4][3][:y] .≈ +101.0)
+    @test all(p12[end-3][3][:y] .≈ +102.0)
+    @test all(p12[end-2][3][:y] .≈ +103.0)
+    @test all(p12[end-1][2][:y] .≈ +104.0)
+    @test all(p12[end-0][2][:y] .≈ +105.0)
 end

test/test_predictive_control.jl

@@ -134,7 +134,8 @@ end
     preparestate!(mpc1, [50, 30])
     updatestate!(mpc1, mpc1.estim.model.uop, [50, 30])
     @test mpc1.estim.x̂0 ≈ [0,0,0,0]
-    preparestate!(mpc1, [50, 30])
+    # do not call preparestate! before moveinput! for the warning:
+    moveinput!(mpc1, [10, 50])
     @test_throws ArgumentError updatestate!(mpc1, [0,0])
 end
 
@@ -348,6 +349,8 @@ end
     model2 = LinModel{Float32}(0.5*ones(1,1), ones(1,1), ones(1,1), zeros(1,0), zeros(1,0), 1.0)
     mpc13  = ExplicitMPC(model2)
     @test isa(mpc13, ExplicitMPC{Float32})
+
+    @test_throws ArgumentError LinMPC(model, Hp=0)
 end
 
 @testset "ExplicitMPC moves and getinfo" begin

test/test_sim_model.jl

@@ -79,6 +79,17 @@ Gss2 = c2d(sys_ss[:,1:2], 0.5Ts, :zoh)
     linmodel12 = LinModel{Float32}(Gss.A, Gss.B, Gss.C, zeros(2, 0), zeros(2, 0), Ts)
     @test isa(linmodel12, LinModel{Float32})
 
+    linmodel13 = LinModel(sys,Ts,i_d=[3])
+    linmodel13 = setname!(linmodel13, 
+        u=["u_c", "u_h"], 
+        y=["y_L", "y_T"], 
+        d=["u_l"],
+        x=["X_1", "X_2", "X_3", "X_4"]
+    )
+    @test all(linmodel13.uname .== ["u_c", "u_h"])
+    @test all(linmodel13.yname .== ["y_L", "y_T"])
+    @test all(linmodel13.dname .== ["u_l"])
+    @test all(linmodel13.xname .== ["X_1", "X_2", "X_3", "X_4"])
 
     @test_throws ErrorException LinModel(sys)
     @test_throws ErrorException LinModel(sys,-Ts)
@@ -101,7 +112,8 @@ end
     @test evaloutput(linmodel1, Float64[]) ≈ linmodel1(Float64[]) ≈ [50,30]
     x = initstate!(linmodel1, [10, 60])
     @test evaloutput(linmodel1) ≈ [50 + 19.0, 30 + 7.4]
-    @test updatestate!(linmodel1, [10, 60]) ≈ x
+    @test preparestate!(linmodel1, [10, 60]) ≈ x
+    @test updatestate!(linmodel1,  [10, 60]) ≈ x
     linmodel2 = LinModel(append(tf(1, [1, 0]), tf(2, [10, 1])), 1.0)
     x = initstate!(linmodel2, [10, 3])
     @test evaloutput(linmodel2) ≈ [0, 6]
@@ -188,13 +200,18 @@ end
     Dd = reshape([0], 1, 1)
     f3(x, u, d) = A*x + Bu*u+ Bd*d
     h3(x, d) = C*x + Dd*d
-    nonlinmodel5 = NonLinModel(f3, h3, 1.0, 1, 2, 1, 1, solver=RungeKutta())
+    solver=RungeKutta()
+    @test string(solver) == 
+        "4th order Runge-Kutta differential equation solver with 1 supersamples."
+    nonlinmodel5 = NonLinModel(f3, h3, 1.0, 1, 2, 1, 1, solver=solver)
     xnext, y = similar(nonlinmodel5.x0), similar(nonlinmodel5.yop)
     nonlinmodel5.f!(xnext, [0; 0], [0], [0])
     @test xnext ≈ zeros(2)
     nonlinmodel5.h!(y, [0; 0], [0])
     @test y ≈ zeros(1)
 
+
+
     function f2!(ẋ, x, u , d)
         mul!(ẋ, A, x)
         mul!(ẋ, Bu, u, 1, 1)