Austin Schuh | b0bfaf8 | 2024-06-19 19:47:23 -0700 | [diff] [blame] | 1 | #ifndef FRC971_CONTROL_LOOPS_RUNGE_KUTTA_HELPERS_H_ |
| 2 | #define FRC971_CONTROL_LOOPS_RUNGE_KUTTA_HELPERS_H_ |
| 3 | |
Austin Schuh | 99f7c6a | 2024-06-25 22:07:44 -0700 | [diff] [blame] | 4 | #include "absl/log/check.h" |
| 5 | #include "absl/log/log.h" |
Austin Schuh | b0bfaf8 | 2024-06-19 19:47:23 -0700 | [diff] [blame] | 6 | #include <Eigen/Dense> |
| 7 | |
| 8 | namespace frc971::control_loops { |
| 9 | |
| 10 | // Returns a reasonable Runge Kutta initial step size. This is translated from |
| 11 | // scipy. |
| 12 | template <typename F, typename T> |
| 13 | double SelectRungeKuttaInitialStep(const F &fn, size_t t0, T y0, T f0, |
| 14 | int error_estimator_order, double rtol, |
| 15 | double atol) { |
| 16 | constexpr int states = y0.rows(); |
| 17 | const Eigen::Matrix<double, states, 1> scale = |
| 18 | atol + (y0.cwiseAbs().matrix() * rtol).array(); |
| 19 | const double sqrt_rows = std::sqrt(static_cast<double>(states)); |
| 20 | const double d0 = (y0.array() / scale.array()).matrix().norm() / sqrt_rows; |
| 21 | const double d1 = (f0.array() / scale.array()).matrix().norm() / sqrt_rows; |
| 22 | double h0; |
| 23 | if (d0 < 1e-5 || d1 < 1e-5) { |
| 24 | h0 = 1e-6; |
| 25 | } else { |
| 26 | h0 = 0.01 * d0 / d1; |
| 27 | } |
| 28 | |
| 29 | const Eigen::Matrix<double, states, 1> y1 = y0 + h0 * f0; |
| 30 | const Eigen::Matrix<double, states, 1> f1 = fn(t0 + h0, y1); |
| 31 | const double d2 = |
| 32 | ((f1 - f0).array() / scale.array()).matrix().norm() / sqrt_rows / h0; |
| 33 | |
| 34 | double h1; |
| 35 | if (d1 <= 1e-15 && d2 <= 1e-15) { |
| 36 | h1 = std::max(1e-6, h0 * 1e-3); |
| 37 | } else { |
| 38 | h1 = std::pow((0.01 / std::max(d1, d2)), |
| 39 | (1.0 / (error_estimator_order + 1.0))); |
| 40 | } |
| 41 | |
| 42 | return std::min(100 * h0, h1); |
| 43 | } |
| 44 | |
| 45 | // Performs a single step of Runge Kutta integration for the adaptive algorithm |
| 46 | // below. This is translated from scipy. |
| 47 | template <size_t N, size_t NStages, size_t Order, typename F> |
| 48 | std::tuple<Eigen::Matrix<double, N, 1>, Eigen::Matrix<double, N, 1>> RKStep( |
| 49 | const F &fn, const double t, const Eigen::Matrix<double, N, 1> &y0, |
| 50 | const Eigen::Matrix<double, N, 1> &f0, const double h, |
| 51 | const Eigen::Matrix<double, NStages, Order> &A, |
| 52 | const Eigen::Matrix<double, 1, NStages> &B, |
| 53 | const Eigen::Matrix<double, 1, NStages> &C, |
| 54 | Eigen::Matrix<double, NStages + 1, N> &K) { |
| 55 | K.template block<N, 1>(0, 0) = f0; |
| 56 | for (size_t s = 1; s < NStages; ++s) { |
| 57 | Eigen::Matrix<double, N, 1> dy = |
| 58 | K.block(0, 0, s, N).transpose() * A.block(s, 0, 1, s).transpose() * h; |
| 59 | K.template block<1, N>(s, 0) = fn(t + C(0, s) * h, y0 + dy).transpose(); |
| 60 | } |
| 61 | |
| 62 | Eigen::Matrix<double, N, 1> y_new = |
| 63 | y0 + h * (K.template block<NStages, N>(0, 0).transpose() * B.transpose()); |
| 64 | Eigen::Matrix<double, N, 1> f_new = fn(t + h, y_new); |
| 65 | |
| 66 | K.template block<1, N>(NStages, 0) = f_new.transpose(); |
| 67 | |
| 68 | return std::make_tuple(y_new, f_new); |
| 69 | } |
| 70 | |
| 71 | } // namespace frc971::control_loops |
| 72 | |
| 73 | #endif // FRC971_CONTROL_LOOPS_RUNGE_KUTTA_HELPERS_H_ |