Add a catapult MPC solver in C++ which runs fast enough

Change-Id: Ieccddb951dce287fbdd74d75916896e97360c50b
Signed-off-by: Austin Schuh <austin.linux@gmail.com>
diff --git a/y2022/control_loops/superstructure/catapult/BUILD b/y2022/control_loops/superstructure/catapult/BUILD
new file mode 100644
index 0000000..289dfde
--- /dev/null
+++ b/y2022/control_loops/superstructure/catapult/BUILD
@@ -0,0 +1,68 @@
+genrule(
+    name = "genrule_catapult",
+    outs = [
+        "catapult_plant.h",
+        "catapult_plant.cc",
+        "integral_catapult_plant.h",
+        "integral_catapult_plant.cc",
+    ],
+    cmd = "$(location //y2022/control_loops/python:catapult) $(OUTS)",
+    target_compatible_with = ["@platforms//os:linux"],
+    tools = [
+        "//y2022/control_loops/python:catapult",
+    ],
+)
+
+cc_library(
+    name = "catapult_plants",
+    srcs = [
+        "catapult_plant.cc",
+        "integral_catapult_plant.cc",
+    ],
+    hdrs = [
+        "catapult_plant.h",
+        "integral_catapult_plant.h",
+    ],
+    visibility = ["//visibility:public"],
+    deps = [
+        "//frc971/control_loops:state_feedback_loop",
+    ],
+)
+
+cc_library(
+    name = "catapult",
+    srcs = [
+        "catapult.cc",
+    ],
+    hdrs = [
+        "catapult.h",
+    ],
+    visibility = ["//visibility:public"],
+    deps = [
+        ":catapult_plants",
+        "//aos:realtime",
+        "//third_party/osqp-cpp",
+    ],
+)
+
+cc_test(
+    name = "catapult_test",
+    srcs = [
+        "catapult_test.cc",
+    ],
+    deps = [
+        ":catapult",
+        "//aos/testing:googletest",
+    ],
+)
+
+cc_binary(
+    name = "catapult_main",
+    srcs = [
+        "catapult_main.cc",
+    ],
+    deps = [
+        ":catapult",
+        "//aos:init",
+    ],
+)
diff --git a/y2022/control_loops/superstructure/catapult/catapult.cc b/y2022/control_loops/superstructure/catapult/catapult.cc
new file mode 100644
index 0000000..243cb2d
--- /dev/null
+++ b/y2022/control_loops/superstructure/catapult/catapult.cc
@@ -0,0 +1,260 @@
+#include "y2022/control_loops/superstructure/catapult/catapult.h"
+
+#include "Eigen/Dense"
+#include "Eigen/Sparse"
+#include "aos/realtime.h"
+#include "aos/time/time.h"
+#include "glog/logging.h"
+#include "osqp++.h"
+#include "osqp.h"
+#include "y2022/control_loops/superstructure/catapult/catapult_plant.h"
+
+namespace y2022 {
+namespace control_loops {
+namespace superstructure {
+namespace catapult {
+namespace chrono = std::chrono;
+
+namespace {
+osqp::OsqpInstance MakeInstance(
+    size_t horizon, Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> P) {
+  osqp::OsqpInstance instance;
+  instance.objective_matrix = P.cast<c_float>().sparseView();
+
+  instance.constraint_matrix =
+      Eigen::SparseMatrix<c_float, Eigen::ColMajor, osqp::c_int>(horizon,
+                                                                 horizon);
+  instance.constraint_matrix.setIdentity();
+
+  instance.lower_bounds =
+      Eigen::Matrix<c_float, Eigen::Dynamic, 1>::Zero(horizon, 1);
+  instance.upper_bounds =
+      Eigen::Matrix<c_float, Eigen::Dynamic, 1>::Ones(horizon, 1) * 12.0;
+  return instance;
+}
+}  // namespace
+
+MPCProblem::MPCProblem(size_t horizon,
+                       Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> P,
+                       Eigen::Matrix<double, Eigen::Dynamic, 1> accel_q,
+                       Eigen::Matrix<double, 2, 2> Af,
+                       Eigen::Matrix<double, Eigen::Dynamic, 2> final_q)
+    : horizon_(horizon),
+      accel_q_(std::move(accel_q.cast<c_float>())),
+      Af_(std::move(Af.cast<c_float>())),
+      final_q_(std::move(final_q.cast<c_float>())),
+      instance_(MakeInstance(horizon, std::move(P))) {
+  instance_.objective_vector =
+      Eigen::Matrix<c_float, Eigen::Dynamic, 1>(horizon, 1);
+  settings_.max_iter = 25;
+  settings_.check_termination = 25;
+  settings_.warm_start = 0;
+  auto status = solver_.Init(instance_, settings_);
+  CHECK(status.ok()) << status;
+}
+
+void MPCProblem::SetState(Eigen::Matrix<c_float, 2, 1> X_initial,
+                          Eigen::Matrix<c_float, 2, 1> X_final) {
+  X_initial_ = X_initial;
+  X_final_ = X_final;
+  instance_.objective_vector =
+      X_initial(1, 0) * accel_q_ + final_q_ * (Af_ * X_initial - X_final);
+
+  auto status = solver_.SetObjectiveVector(instance_.objective_vector);
+  CHECK(status.ok()) << status;
+}
+
+bool MPCProblem::Solve() {
+  const aos::monotonic_clock::time_point start_time =
+      aos::monotonic_clock::now();
+  osqp::OsqpExitCode exit_code = solver_.Solve();
+  const aos::monotonic_clock::time_point end_time = aos::monotonic_clock::now();
+  VLOG(1) << "OSQP solved in "
+          << std::chrono::duration<double>(end_time - start_time).count();
+  solve_time_ = std::chrono::duration<double>(end_time - start_time).count();
+  // TODO(austin): Dump the exit codes out as an enum for logging.
+  //
+  // TODO(austin): The dual problem doesn't appear to be converging on all
+  // problems.  Are we phrasing something wrong?
+
+  // TODO(austin): Set a time limit so we can't run forever, and signal back
+  // when we hit our limit.
+  return exit_code == osqp::OsqpExitCode::kOptimal;
+}
+
+void MPCProblem::WarmStart(const MPCProblem &p) {
+  CHECK_GE(p.horizon(), horizon())
+      << ": Can only copy a bigger problem's solution into a smaller problem.";
+  auto status = solver_.SetPrimalWarmStart(p.solver_.primal_solution().block(
+      p.horizon() - horizon(), 0, horizon(), 1));
+  CHECK(status.ok()) << status;
+  status = solver_.SetDualWarmStart(p.solver_.dual_solution().block(
+      p.horizon() - horizon(), 0, horizon(), 1));
+  CHECK(status.ok()) << status;
+}
+
+CatapultProblemGenerator::CatapultProblemGenerator(size_t horizon)
+    : plant_(MakeCatapultPlant()),
+      horizon_(horizon),
+      Q_final_(
+          (Eigen::DiagonalMatrix<double, 2>().diagonal() << 10000.0, 10000.0)
+              .finished()),
+      As_(MakeAs()),
+      Bs_(MakeBs()),
+      m_(Makem()),
+      M_(MakeM()),
+      W_(MakeW()),
+      w_(Makew()),
+      Pi_(MakePi()),
+      WM_(W_ * M_),
+      Wmpw_(W_ * m_ + w_) {}
+
+std::unique_ptr<MPCProblem> CatapultProblemGenerator::MakeProblem(
+    size_t horizon) {
+  return std::make_unique<MPCProblem>(
+      horizon, P(horizon), accel_q(horizon), Af(horizon),
+      (2.0 * Q_final_ * Bf(horizon)).transpose());
+}
+
+const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>
+CatapultProblemGenerator::P(size_t horizon) {
+  CHECK_GT(horizon, 0u);
+  CHECK_LE(horizon, horizon_);
+  return 2.0 * (WM_.block(0, 0, horizon, horizon).transpose() * Pi(horizon) *
+                    WM_.block(0, 0, horizon, horizon) +
+                Bf(horizon).transpose() * Q_final_ * Bf(horizon));
+}
+
+const Eigen::Matrix<double, Eigen::Dynamic, 1> CatapultProblemGenerator::q(
+    size_t horizon, Eigen::Matrix<c_float, 2, 1> X_initial,
+    Eigen::Matrix<c_float, 2, 1> X_final) {
+  CHECK_GT(horizon, 0u);
+  CHECK_LE(horizon, horizon_);
+  return 2.0 * X_initial(1, 0) * accel_q(horizon) +
+         2.0 * ((Af(horizon) * X_initial - X_final).transpose() * Q_final_ *
+                Bf(horizon))
+                   .transpose();
+}
+
+const Eigen::Matrix<double, Eigen::Dynamic, 1>
+CatapultProblemGenerator::accel_q(size_t horizon) {
+  return 2.0 * ((Wmpw_.block(0, 0, horizon, 1)).transpose() * Pi(horizon) *
+                WM_.block(0, 0, horizon, horizon))
+                   .transpose();
+}
+
+const Eigen::Matrix<double, 2, 2> CatapultProblemGenerator::Af(size_t horizon) {
+  CHECK_GT(horizon, 0u);
+  CHECK_LE(horizon, horizon_);
+  return As_.block<2, 2>(2 * (horizon - 1), 0);
+}
+
+const Eigen::Matrix<double, 2, Eigen::Dynamic> CatapultProblemGenerator::Bf(
+    size_t horizon) {
+  CHECK_GT(horizon, 0u);
+  CHECK_LE(horizon, horizon_);
+  return Bs_.block(2 * (horizon - 1), 0, 2, horizon);
+}
+
+const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>
+CatapultProblemGenerator::Pi(size_t horizon) {
+  CHECK_GT(horizon, 0u);
+  CHECK_LE(horizon, horizon_);
+  return Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>(Pi_).block(
+      horizon_ - horizon, horizon_ - horizon, horizon, horizon);
+}
+
+Eigen::Matrix<double, Eigen::Dynamic, 2> CatapultProblemGenerator::MakeAs() {
+  Eigen::Matrix<double, Eigen::Dynamic, 2> As =
+      Eigen::Matrix<double, Eigen::Dynamic, 2>::Zero(horizon_ * 2, 2);
+  for (size_t i = 0; i < horizon_; ++i) {
+    if (i == 0) {
+      As.block<2, 2>(0, 0) = plant_.A();
+    } else {
+      As.block<2, 2>(i * 2, 0) = plant_.A() * As.block<2, 2>((i - 1) * 2, 0);
+    }
+  }
+  return As;
+}
+
+Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>
+CatapultProblemGenerator::MakeBs() {
+  Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> Bs =
+      Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>::Zero(horizon_ * 2,
+                                                                  horizon_);
+  for (size_t i = 0; i < horizon_; ++i) {
+    for (size_t j = 0; j < i + 1; ++j) {
+      if (i == j) {
+        Bs.block<2, 1>(i * 2, j) = plant_.B();
+      } else {
+        Bs.block<2, 1>(i * 2, j) =
+            As_.block<2, 2>((i - j - 1) * 2, 0) * plant_.B();
+      }
+    }
+  }
+  return Bs;
+}
+
+Eigen::Matrix<double, Eigen::Dynamic, 1> CatapultProblemGenerator::Makem() {
+  Eigen::Matrix<double, Eigen::Dynamic, 1> m =
+      Eigen::Matrix<double, Eigen::Dynamic, 1>::Zero(horizon_, 1);
+  for (size_t i = 0; i < horizon_; ++i) {
+    m(i, 0) = As_(1 + 2 * i, 1);
+  }
+  return m;
+}
+
+Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>
+CatapultProblemGenerator::MakeM() {
+  Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> M =
+      Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>::Zero(horizon_,
+                                                                  horizon_);
+  for (size_t i = 0; i < horizon_; ++i) {
+    for (size_t j = 0; j < horizon_; ++j) {
+      M(i, j) = Bs_(2 * i + 1, j);
+    }
+  }
+  return M;
+}
+
+Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>
+CatapultProblemGenerator::MakeW() {
+  Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> W =
+      Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>::Identity(horizon_,
+                                                                      horizon_);
+  for (size_t i = 0; i < horizon_ - 1; ++i) {
+    W(i + 1, i) = -1.0;
+  }
+  W /= std::chrono::duration<double>(plant_.dt()).count();
+  return W;
+}
+
+Eigen::Matrix<double, Eigen::Dynamic, 1> CatapultProblemGenerator::Makew() {
+  Eigen::Matrix<double, Eigen::Dynamic, 1> w =
+      Eigen::Matrix<double, Eigen::Dynamic, 1>::Zero(horizon_, 1);
+  w(0, 0) = -1.0 / std::chrono::duration<double>(plant_.dt()).count();
+  return w;
+}
+
+Eigen::DiagonalMatrix<double, Eigen::Dynamic>
+CatapultProblemGenerator::MakePi() {
+  Eigen::DiagonalMatrix<double, Eigen::Dynamic> Pi(horizon_);
+  for (size_t i = 0; i < horizon_; ++i) {
+    Pi.diagonal()(i) =
+        std::pow(0.01, 2.0) +
+        std::pow(0.02 * std::max(0.0, (20 - ((int)horizon_ - (int)i)) / 20.),
+                 2.0);
+  }
+  return Pi;
+}
+
+Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>
+CatapultProblemGenerator::MakeP() {
+  return 2.0 * (M_.transpose() * W_.transpose() * Pi_ * W_ * M_ +
+                Bf(horizon_).transpose() * Q_final_ * Bf(horizon_));
+}
+
+}  // namespace catapult
+}  // namespace superstructure
+}  // namespace control_loops
+}  // namespace y2022
diff --git a/y2022/control_loops/superstructure/catapult/catapult.h b/y2022/control_loops/superstructure/catapult/catapult.h
new file mode 100644
index 0000000..85e9242
--- /dev/null
+++ b/y2022/control_loops/superstructure/catapult/catapult.h
@@ -0,0 +1,132 @@
+#ifndef Y2022_CONTROL_LOOPS_SUPERSTRUCTURE_CATAPULT_CATAPULT_H_
+#define Y2022_CONTROL_LOOPS_SUPERSTRUCTURE_CATAPULT_CATAPULT_H_
+
+#include "Eigen/Dense"
+#include "frc971/control_loops/state_feedback_loop.h"
+#include "glog/logging.h"
+#include "osqp++.h"
+
+namespace y2022 {
+namespace control_loops {
+namespace superstructure {
+namespace catapult {
+
+// MPC problem for a specified horizon.  This contains all the state for the
+// solver, setters to modify the current and target state, and a way to fetch
+// the solution.
+class MPCProblem {
+ public:
+  MPCProblem(size_t horizon,
+             Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> P,
+             Eigen::Matrix<double, Eigen::Dynamic, 1> accel_q,
+             Eigen::Matrix<double, 2, 2> Af,
+             Eigen::Matrix<double, Eigen::Dynamic, 2> final_q);
+
+  MPCProblem(MPCProblem const &) = delete;
+  void operator=(MPCProblem const &x) = delete;
+
+  // Sets the current and final state.  This keeps the problem in tact and
+  // doesn't recreate it, so it will be fast.
+  void SetState(Eigen::Matrix<double, 2, 1> X_initial,
+                Eigen::Matrix<double, 2, 1> X_final);
+
+  // Solves our problem.
+  bool Solve();
+
+  double solve_time() const { return solve_time_; }
+
+  // Returns the solution that the solver found when Solve was last called.
+  double U(size_t i) const { return solver_.primal_solution()(i); }
+
+  // Returns the number of U's to be solved.
+  size_t horizon() const { return horizon_; }
+
+  // Warm starts the optimizer with the provided solution to make it solve
+  // faster.
+  void WarmStart(const MPCProblem &p);
+
+ private:
+  // The number of u's to solve for.
+  const size_t horizon_;
+
+  // The problem statement variables needed by SetState to update q.
+  const Eigen::Matrix<double, Eigen::Dynamic, 1> accel_q_;
+  const Eigen::Matrix<double, 2, 2> Af_;
+  const Eigen::Matrix<double, Eigen::Dynamic, 2> final_q_;
+
+  Eigen::Matrix<double, 2, 1> X_initial_;
+  Eigen::Matrix<double, 2, 1> X_final_;
+
+  // Solver state.
+  osqp::OsqpInstance instance_;
+  osqp::OsqpSolver solver_;
+  osqp::OsqpSettings settings_;
+
+  double solve_time_ = 0;
+};
+
+// Decently efficient problem generator for multiple horizons given a max
+// horizon to solve for.
+//
+// The math is documented in mpc.tex
+class CatapultProblemGenerator {
+ public:
+  // Builds a problem generator for the specified max horizon and caches a lot
+  // of the state.
+  CatapultProblemGenerator(size_t horizon);
+
+  // Returns the maximum horizon.
+  size_t horizon() const { return horizon_; }
+
+  // Makes a problem for the specificed horizon.
+  std::unique_ptr<MPCProblem> MakeProblem(size_t horizon);
+
+  // Returns the P and Q matrices for the problem statement.
+  //   cost = 0.5 X.T P X + q.T X
+  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> P(size_t horizon);
+  const Eigen::Matrix<double, Eigen::Dynamic, 1> q(
+      size_t horizon, Eigen::Matrix<double, 2, 1> X_initial,
+      Eigen::Matrix<double, 2, 1> X_final);
+
+ private:
+  const Eigen::Matrix<double, Eigen::Dynamic, 1> accel_q(size_t horizon);
+
+  const Eigen::Matrix<double, 2, 2> Af(size_t horizon);
+  const Eigen::Matrix<double, 2, Eigen::Dynamic> Bf(size_t horizon);
+  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> Pi(
+      size_t horizon);
+
+  // These functions are used in the constructor to build up the matrices below.
+  Eigen::Matrix<double, Eigen::Dynamic, 2> MakeAs();
+  Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> MakeBs();
+  Eigen::Matrix<double, Eigen::Dynamic, 1> Makem();
+  Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> MakeM();
+  Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> MakeW();
+  Eigen::Matrix<double, Eigen::Dynamic, 1> Makew();
+  Eigen::DiagonalMatrix<double, Eigen::Dynamic> MakePi();
+  Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> MakeP();
+
+  const StateFeedbackPlant<2, 1, 1> plant_;
+  const size_t horizon_;
+
+  const Eigen::DiagonalMatrix<double, 2> Q_final_;
+
+  const Eigen::Matrix<double, Eigen::Dynamic, 2> As_;
+  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> Bs_;
+  const Eigen::Matrix<double, Eigen::Dynamic, 1> m_;
+  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> M_;
+
+  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> W_;
+  const Eigen::Matrix<double, Eigen::Dynamic, 1> w_;
+  const Eigen::DiagonalMatrix<double, Eigen::Dynamic> Pi_;
+
+  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> WM_;
+  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> Wmpw_;
+};
+
+}  // namespace catapult
+}  // namespace superstructure
+}  // namespace control_loops
+}  // namespace y2022
+
+#endif  // Y2022_CONTROL_LOOPS_SUPERSTRUCTURE_CATAPULT_CATAPULT_H_
diff --git a/y2022/control_loops/superstructure/catapult/catapult_main.cc b/y2022/control_loops/superstructure/catapult/catapult_main.cc
new file mode 100644
index 0000000..53d357a
--- /dev/null
+++ b/y2022/control_loops/superstructure/catapult/catapult_main.cc
@@ -0,0 +1,126 @@
+#include "aos/init.h"
+#include "aos/realtime.h"
+#include "aos/time/time.h"
+#include "y2022/control_loops/superstructure/catapult/catapult.h"
+#include "y2022/control_loops/superstructure/catapult/catapult_plant.h"
+
+namespace y2022 {
+namespace control_loops {
+namespace superstructure {
+namespace catapult {
+namespace chrono = std::chrono;
+
+void OSQPSolve() {
+  Eigen::Matrix<double, 2, 1> X_initial(0.0, 0.0);
+  Eigen::Matrix<double, 2, 1> X_final(2.0, 25.0);
+
+  LOG(INFO) << "Starting a dynamic OSQP solve";
+  CatapultProblemGenerator g(35);
+  const StateFeedbackPlant<2, 1, 1> plant = MakeCatapultPlant();
+
+  constexpr int kHorizon = 35;
+
+  // TODO(austin): This is a good unit test!  Make sure computing the problem
+  // different ways comes out the same.
+  {
+    CatapultProblemGenerator g2(10);
+    constexpr int kTestHorizon = 10;
+    CHECK(g2.P(kTestHorizon) == g.P(kTestHorizon))
+        << g2.P(kTestHorizon) - g.P(kTestHorizon);
+    CHECK(g2.q(kTestHorizon, X_initial, X_final) ==
+          g.q(kTestHorizon, X_initial, X_final))
+        << g2.q(kTestHorizon, X_initial, X_final) -
+               g.q(kTestHorizon, X_initial, X_final);
+  }
+
+  osqp::OsqpInstance instance;
+  instance.objective_matrix = g.P(kHorizon).sparseView();
+
+  instance.objective_vector = g.q(kHorizon, X_initial, X_final);
+
+  instance.constraint_matrix =
+      Eigen::SparseMatrix<double, Eigen::ColMajor, osqp::c_int>(kHorizon,
+                                                                kHorizon);
+  instance.constraint_matrix.setIdentity();
+
+  instance.lower_bounds =
+      Eigen::Matrix<double, Eigen::Dynamic, 1>::Zero(kHorizon, 1);
+  instance.upper_bounds =
+      Eigen::Matrix<double, Eigen::Dynamic, 1>::Ones(kHorizon, 1) * 12.0;
+
+  osqp::OsqpSolver solver;
+  osqp::OsqpSettings settings;
+  // Edit settings if appropriate.
+  auto status = solver.Init(instance, settings);
+  CHECK(status.ok()) << status;
+
+  aos::LockAllMemory();
+  aos::ExpandStackSize();
+  aos::SetCurrentThreadRealtimePriority(60);
+
+  for (int i = 0; i < 10; ++i) {
+    const aos::monotonic_clock::time_point start_time =
+        aos::monotonic_clock::now();
+    osqp::OsqpExitCode exit_code = solver.Solve();
+    const aos::monotonic_clock::time_point end_time =
+        aos::monotonic_clock::now();
+    LOG(INFO) << "OSQP solved in "
+              << chrono::duration<double>(end_time - start_time).count();
+    CHECK(exit_code == osqp::OsqpExitCode::kOptimal);
+  }
+
+  double optimal_objective = solver.objective_value();
+  Eigen::Matrix<double, Eigen::Dynamic, 1> optimal_solution =
+      solver.primal_solution();
+
+  LOG(INFO) << "Cost: " << optimal_objective;
+  LOG(INFO) << "U: " << optimal_solution;
+
+  std::vector<std::unique_ptr<MPCProblem>> problems;
+  for (size_t i = g.horizon(); i > 0; --i) {
+    LOG(INFO) << "Made problem " << i;
+    problems.emplace_back(g.MakeProblem(i));
+  }
+
+  std::unique_ptr<MPCProblem> p = g.MakeProblem(kHorizon);
+
+  p->SetState(X_initial, X_final);
+  p->Solve();
+  p->Solve();
+  p->Solve();
+  p->Solve();
+
+  Eigen::Vector2d X = X_initial;
+  for (size_t i = 0; i < g.horizon(); ++i) {
+    problems[i]->SetState(X, X_final);
+    if (i != 0) {
+      problems[i]->WarmStart(*problems[i - 1]);
+    }
+
+    problems[i]->Solve();
+    X = plant.A() * X + plant.B() * problems[i]->U(0);
+
+    LOG(INFO) << "Dynamic u(" << i << "): " << problems[i]->U(0) << " -> "
+              << X.transpose();
+  }
+
+  aos::UnsetCurrentThreadRealtimePriority();
+}
+
+int Main(int /*argc*/, char ** /*argv*/) {
+  OSQPSolve();
+
+  gflags::ShutDownCommandLineFlags();
+  return 0;
+}
+
+}  // namespace catapult
+}  // namespace superstructure
+}  // namespace control_loops
+}  // namespace y2022
+
+int main(int argc, char **argv) {
+  ::aos::InitGoogle(&argc, &argv);
+
+  return y2022::control_loops::superstructure::catapult::Main(argc, argv);
+}
diff --git a/y2022/control_loops/superstructure/catapult/catapult_test.cc b/y2022/control_loops/superstructure/catapult/catapult_test.cc
new file mode 100644
index 0000000..b37c3e0
--- /dev/null
+++ b/y2022/control_loops/superstructure/catapult/catapult_test.cc
@@ -0,0 +1,62 @@
+#include "y2022/control_loops/superstructure/catapult/catapult.h"
+
+#include "glog/logging.h"
+#include "gtest/gtest.h"
+#include "y2022/control_loops/superstructure/catapult/catapult_plant.h"
+
+namespace y2022 {
+namespace control_loops {
+namespace superstructure {
+namespace catapult {
+namespace testing {
+
+// Tests that computing P and q with 2 different horizons comes out the same.
+TEST(MPCTest, HorizonTest) {
+  Eigen::Matrix<double, 2, 1> X_initial(0.0, 0.0);
+  Eigen::Matrix<double, 2, 1> X_final(2.0, 25.0);
+
+  CatapultProblemGenerator g(35);
+
+  CatapultProblemGenerator g2(10);
+  constexpr int kTestHorizon = 10;
+  EXPECT_TRUE(g2.P(kTestHorizon) == g.P(kTestHorizon))
+      << g2.P(kTestHorizon) - g.P(kTestHorizon);
+  EXPECT_TRUE(g2.q(kTestHorizon, X_initial, X_final) ==
+              g.q(kTestHorizon, X_initial, X_final))
+      << g2.q(kTestHorizon, X_initial, X_final) -
+             g.q(kTestHorizon, X_initial, X_final);
+}
+
+// Tests that we can get to the destination.
+TEST(MPCTest, NearGoal) {
+  Eigen::Matrix<double, 2, 1> X_initial(0.0, 0.0);
+  Eigen::Matrix<double, 2, 1> X_final(2.0, 25.0);
+
+  CatapultProblemGenerator g(35);
+  const StateFeedbackPlant<2, 1, 1> plant = MakeCatapultPlant();
+
+  std::vector<std::unique_ptr<MPCProblem>> problems;
+  for (size_t i = g.horizon(); i > 0; --i) {
+    problems.emplace_back(g.MakeProblem(i));
+  }
+
+  Eigen::Vector2d X = X_initial;
+  for (size_t i = 0; i < g.horizon(); ++i) {
+    problems[i]->SetState(X, X_final);
+    if (i != 0) {
+      problems[i]->WarmStart(*problems[i - 1]);
+    }
+
+    problems[i]->Solve();
+    X = plant.A() * X + plant.B() * problems[i]->U(0);
+  }
+
+  EXPECT_NEAR(X_final.x(), X.x(), 1e-2);
+  EXPECT_NEAR(X_final.y(), X.y(), 1e-2);
+}
+
+}  // namespace testing
+}  // namespace catapult
+}  // namespace superstructure
+}  // namespace control_loops
+}  // namespace y2022