y2017/control_loops/python/column.py

#!/usr/bin/python3

from aos.util.trapezoid_profile import TrapezoidProfile
from frc971.control_loops.python import control_loop
from frc971.control_loops.python import controls
from y2017.control_loops.python import turret
from y2017.control_loops.python import indexer
import numpy
import sys
from matplotlib import pylab
import gflags
import glog

FLAGS = gflags.FLAGS

try:
    gflags.DEFINE_bool('plot', False, 'If true, plot the loop response.')
except gflags.DuplicateFlagError:
    pass


class ColumnController(control_loop.ControlLoop):

    def __init__(self, name='Column'):
        super(ColumnController, self).__init__(name)
        self.turret = turret.Turret(name + 'Turret')
        self.indexer = indexer.Indexer(name + 'Indexer')

        # Control loop time step
        self.dt = 0.005

        # State is [position_indexer,
        #           velocity_indexer,
        #           position_shooter,
        #           velocity_shooter]
        # Input is [volts_indexer, volts_shooter]
        self.A_continuous = numpy.matrix(numpy.zeros((3, 3)))
        self.B_continuous = numpy.matrix(numpy.zeros((3, 2)))

        self.A_continuous[0, 0] = -(self.indexer.Kt / self.indexer.Kv /
                                    (self.indexer.J * self.indexer.resistance *
                                     self.indexer.G * self.indexer.G) +
                                    self.turret.Kt / self.turret.Kv /
                                    (self.indexer.J * self.turret.resistance *
                                     self.turret.G * self.turret.G))
        self.A_continuous[0, 2] = self.turret.Kt / self.turret.Kv / (
            self.indexer.J * self.turret.resistance * self.turret.G *
            self.turret.G)
        self.B_continuous[0, 0] = self.indexer.Kt / (
            self.indexer.J * self.indexer.resistance * self.indexer.G)
        self.B_continuous[0, 1] = -self.turret.Kt / (
            self.indexer.J * self.turret.resistance * self.turret.G)

        self.A_continuous[1, 2] = 1

        self.A_continuous[2, 0] = self.turret.Kt / self.turret.Kv / (
            self.turret.J * self.turret.resistance * self.turret.G *
            self.turret.G)
        self.A_continuous[2, 2] = -self.turret.Kt / self.turret.Kv / (
            self.turret.J * self.turret.resistance * self.turret.G *
            self.turret.G)

        self.B_continuous[2, 1] = self.turret.Kt / (
            self.turret.J * self.turret.resistance * self.turret.G)

        self.C = numpy.matrix([[1, 0, 0], [0, 1, 0]])
        self.D = numpy.matrix([[0, 0], [0, 0]])

        self.A, self.B = self.ContinuousToDiscrete(self.A_continuous,
                                                   self.B_continuous, self.dt)

        q_indexer_vel = 13.0
        q_pos = 0.05
        q_vel = 0.8
        self.Q = numpy.matrix([[(1.0 / (q_indexer_vel**2.0)), 0.0, 0.0],
                               [0.0, (1.0 / (q_pos**2.0)), 0.0],
                               [0.0, 0.0, (1.0 / (q_vel**2.0))]])

        self.R = numpy.matrix([[(1.0 / (12.0**2.0)), 0.0],
                               [0.0, (1.0 / (12.0**2.0))]])
        self.K = controls.dlqr(self.A, self.B, self.Q, self.R)

        glog.debug('Controller poles are ' +
                   repr(numpy.linalg.eig(self.A - self.B * self.K)[0]))

        q_vel_indexer_ff = 0.000005
        q_pos_ff = 0.0000005
        q_vel_ff = 0.00008
        self.Qff = numpy.matrix([[(1.0 / (q_vel_indexer_ff**2.0)), 0.0, 0.0],
                                 [0.0, (1.0 / (q_pos_ff**2.0)), 0.0],
                                 [0.0, 0.0, (1.0 / (q_vel_ff**2.0))]])

        self.Kff = controls.TwoStateFeedForwards(self.B, self.Qff)

        self.U_max = numpy.matrix([[12.0], [12.0]])
        self.U_min = numpy.matrix([[-12.0], [-12.0]])

        self.InitializeState()


class Column(ColumnController):

    def __init__(self, name='Column', disable_indexer=False):
        super(Column, self).__init__(name)
        A_continuous = numpy.matrix(numpy.zeros((4, 4)))
        B_continuous = numpy.matrix(numpy.zeros((4, 2)))

        A_continuous[0, 1] = 1
        A_continuous[1:, 1:] = self.A_continuous
        B_continuous[1:, :] = self.B_continuous

        self.A_continuous = A_continuous
        self.B_continuous = B_continuous

        self.A, self.B = self.ContinuousToDiscrete(self.A_continuous,
                                                   self.B_continuous, self.dt)

        self.C = numpy.matrix([[1, 0, 0, 0], [-1, 0, 1, 0]])
        self.D = numpy.matrix([[0, 0], [0, 0]])

        orig_K = self.K
        self.K = numpy.matrix(numpy.zeros((2, 4)))
        self.K[:, 1:] = orig_K

        glog.debug('K is ' + repr(self.K))
        # TODO(austin): Do we want to damp velocity out or not when disabled?
        #if disable_indexer:
        #  self.K[0, 1] = 0.0
        #  self.K[1, 1] = 0.0

        orig_Kff = self.Kff
        self.Kff = numpy.matrix(numpy.zeros((2, 4)))
        self.Kff[:, 1:] = orig_Kff

        q_pos = 0.12
        q_vel = 2.00
        self.Q = numpy.matrix([[(q_pos**2.0), 0.0, 0.0, 0.0],
                               [0.0, (q_vel**2.0), 0.0, 0.0],
                               [0.0, 0.0, (q_pos**2.0), 0.0],
                               [0.0, 0.0, 0.0, (q_vel**2.0)]])

        r_pos = 0.05
        self.R = numpy.matrix([[(r_pos**2.0), 0.0], [0.0, (r_pos**2.0)]])

        self.KalmanGain, self.Q_steady = controls.kalman(A=self.A,
                                                         B=self.B,
                                                         C=self.C,
                                                         Q=self.Q,
                                                         R=self.R)
        self.L = self.A * self.KalmanGain

        self.InitializeState()


class IntegralColumn(Column):

    def __init__(self,
                 name='IntegralColumn',
                 voltage_error_noise=None,
                 disable_indexer=False):
        super(IntegralColumn, self).__init__(name)

        A_continuous = numpy.matrix(numpy.zeros((6, 6)))
        A_continuous[0:4, 0:4] = self.A_continuous
        A_continuous[0:4:, 4:6] = self.B_continuous

        B_continuous = numpy.matrix(numpy.zeros((6, 2)))
        B_continuous[0:4, :] = self.B_continuous

        self.A_continuous = A_continuous
        self.B_continuous = B_continuous

        self.A, self.B = self.ContinuousToDiscrete(self.A_continuous,
                                                   self.B_continuous, self.dt)

        C = numpy.matrix(numpy.zeros((2, 6)))
        C[0:2, 0:4] = self.C
        self.C = C

        self.D = numpy.matrix([[0, 0], [0, 0]])

        orig_K = self.K
        self.K = numpy.matrix(numpy.zeros((2, 6)))
        self.K[:, 0:4] = orig_K

        # TODO(austin): I'm not certain this is ideal.  If someone spins the bottom
        # at a constant rate, we'll learn a voltage offset.  That should translate
        # directly to a voltage on the turret to hold it steady.  I'm also not
        # convinced we care that much.  If the indexer is off, it'll stop rather
        # quickly anyways, so this is mostly a moot point.
        if not disable_indexer:
            self.K[0, 4] = 1
        self.K[1, 5] = 1

        orig_Kff = self.Kff
        self.Kff = numpy.matrix(numpy.zeros((2, 6)))
        self.Kff[:, 0:4] = orig_Kff

        q_pos = 0.40
        q_vel = 2.00
        q_voltage = 8.0
        if voltage_error_noise is not None:
            q_voltage = voltage_error_noise

        self.Q = numpy.matrix([[(q_pos**2.0), 0.0, 0.0, 0.0, 0.0, 0.0],
                               [0.0, (q_vel**2.0), 0.0, 0.0, 0.0, 0.0],
                               [0.0, 0.0, (q_pos**2.0), 0.0, 0.0, 0.0],
                               [0.0, 0.0, 0.0, (q_vel**2.0), 0.0, 0.0],
                               [0.0, 0.0, 0.0, 0.0, (q_voltage**2.0), 0.0],
                               [0.0, 0.0, 0.0, 0.0, 0.0, (q_voltage**2.0)]])

        r_pos = 0.05
        self.R = numpy.matrix([[(r_pos**2.0), 0.0], [0.0, (r_pos**2.0)]])

        self.KalmanGain, self.Q_steady = controls.kalman(A=self.A,
                                                         B=self.B,
                                                         C=self.C,
                                                         Q=self.Q,
                                                         R=self.R)
        self.L = self.A * self.KalmanGain

        self.InitializeState()


class ScenarioPlotter(object):

    def __init__(self):
        # Various lists for graphing things.
        self.t = []
        self.xi = []
        self.xt = []
        self.vi = []
        self.vt = []
        self.ai = []
        self.at = []
        self.x_hat = []
        self.ui = []
        self.ut = []
        self.ui_fb = []
        self.ut_fb = []
        self.offseti = []
        self.offsett = []
        self.turret_error = []

    def run_test(self,
                 column,
                 end_goal,
                 controller_column,
                 observer_column=None,
                 iterations=200):
        """Runs the column plant with an initial condition and goal.

        Args:
            column: column object to use.
            end_goal: end_goal state.
            controller_column: Intake object to get K from, or None if we should
                use column.
            observer_column: Intake object to use for the observer, or None if we should
                use the actual state.
            iterations: Number of timesteps to run the model for.
        """

        if controller_column is None:
            controller_column = column

        vbat = 12.0

        if self.t:
            initial_t = self.t[-1] + column.dt
        else:
            initial_t = 0

        goal = numpy.concatenate((column.X, numpy.matrix(numpy.zeros((2, 1)))),
                                 axis=0)

        profile = TrapezoidProfile(column.dt)
        profile.set_maximum_acceleration(10.0)
        profile.set_maximum_velocity(3.0)
        profile.SetGoal(goal[2, 0])

        U_last = numpy.matrix(numpy.zeros((2, 1)))
        for i in range(iterations):
            observer_column.Y = column.Y
            observer_column.CorrectObserver(U_last)

            self.offseti.append(observer_column.X_hat[4, 0])
            self.offsett.append(observer_column.X_hat[5, 0])
            self.x_hat.append(observer_column.X_hat[0, 0])

            next_goal = numpy.concatenate(
                (end_goal[0:2, :],
                 profile.Update(end_goal[2, 0],
                                end_goal[3, 0]), end_goal[4:6, :]),
                axis=0)

            ff_U = controller_column.Kff * (next_goal -
                                            observer_column.A * goal)
            fb_U = controller_column.K * (goal - observer_column.X_hat)
            self.turret_error.append((goal[2, 0] - column.X[2, 0]) * 100.0)
            self.ui_fb.append(fb_U[0, 0])
            self.ut_fb.append(fb_U[1, 0])

            U_uncapped = ff_U + fb_U

            U = U_uncapped.copy()
            U[0, 0] = numpy.clip(U[0, 0], -vbat, vbat)
            U[1, 0] = numpy.clip(U[1, 0], -vbat, vbat)
            self.xi.append(column.X[0, 0])
            self.xt.append(column.X[2, 0])

            if self.vi:
                last_vi = self.vi[-1]
            else:
                last_vi = 0
            if self.vt:
                last_vt = self.vt[-1]
            else:
                last_vt = 0

            self.vi.append(column.X[1, 0])
            self.vt.append(column.X[3, 0])
            self.ai.append((self.vi[-1] - last_vi) / column.dt)
            self.at.append((self.vt[-1] - last_vt) / column.dt)

            offset = 0.0
            if i > 100:
                offset = 1.0
            column.Update(U + numpy.matrix([[0.0], [offset]]))

            observer_column.PredictObserver(U)

            self.t.append(initial_t + i * column.dt)
            self.ui.append(U[0, 0])
            self.ut.append(U[1, 0])

            ff_U -= U_uncapped - U
            goal = controller_column.A * goal + controller_column.B * ff_U

            if U[1, 0] != U_uncapped[1, 0]:
                profile.MoveCurrentState(
                    numpy.matrix([[goal[2, 0]], [goal[3, 0]]]))

        glog.debug('Time: %f', self.t[-1])
        glog.debug('goal_error %s', repr((end_goal - goal).T))
        glog.debug('error %s', repr((observer_column.X_hat - end_goal).T))

    def Plot(self):
        pylab.subplot(3, 1, 1)
        pylab.plot(self.t, self.xi, label='x_indexer')
        pylab.plot(self.t, self.xt, label='x_turret')
        pylab.plot(self.t, self.x_hat, label='x_hat')
        pylab.plot(self.t, self.turret_error, label='turret_error * 100')
        pylab.legend()

        pylab.subplot(3, 1, 2)
        pylab.plot(self.t, self.ui, label='u_indexer')
        pylab.plot(self.t, self.ui_fb, label='u_indexer_fb')
        pylab.plot(self.t, self.ut, label='u_turret')
        pylab.plot(self.t, self.ut_fb, label='u_turret_fb')
        pylab.plot(self.t, self.offseti, label='voltage_offset_indexer')
        pylab.plot(self.t, self.offsett, label='voltage_offset_turret')
        pylab.legend()

        pylab.subplot(3, 1, 3)
        pylab.plot(self.t, self.ai, label='a_indexer')
        pylab.plot(self.t, self.at, label='a_turret')
        pylab.plot(self.t, self.vi, label='v_indexer')
        pylab.plot(self.t, self.vt, label='v_turret')
        pylab.legend()

        pylab.show()


def main(argv):
    scenario_plotter = ScenarioPlotter()

    column = Column()
    column_controller = IntegralColumn()
    observer_column = IntegralColumn()

    initial_X = numpy.matrix([[0.0], [0.0], [0.0], [0.0]])
    R = numpy.matrix([[0.0], [10.0], [5.0], [0.0], [0.0], [0.0]])
    scenario_plotter.run_test(column,
                              end_goal=R,
                              controller_column=column_controller,
                              observer_column=observer_column,
                              iterations=400)

    if FLAGS.plot:
        scenario_plotter.Plot()

    if len(argv) != 7:
        glog.fatal('Expected .h file name and .cc file names')
    else:
        namespaces = ['y2017', 'control_loops', 'superstructure', 'column']
        column = Column('Column')
        loop_writer = control_loop.ControlLoopWriter('Column', [column],
                                                     namespaces=namespaces)
        loop_writer.AddConstant(
            control_loop.Constant('kIndexerFreeSpeed', '%f',
                                  column.indexer.free_speed))
        loop_writer.AddConstant(
            control_loop.Constant('kIndexerOutputRatio', '%f',
                                  column.indexer.G))
        loop_writer.AddConstant(
            control_loop.Constant('kTurretFreeSpeed', '%f',
                                  column.turret.free_speed))
        loop_writer.AddConstant(
            control_loop.Constant('kTurretOutputRatio', '%f', column.turret.G))
        loop_writer.Write(argv[1], argv[2])

        # IntegralColumn controller 1 will disable the indexer.
        integral_column = IntegralColumn('IntegralColumn')
        disabled_integral_column = IntegralColumn('DisabledIntegralColumn',
                                                  disable_indexer=True)
        integral_loop_writer = control_loop.ControlLoopWriter(
            'IntegralColumn', [integral_column, disabled_integral_column],
            namespaces=namespaces)
        integral_loop_writer.Write(argv[3], argv[4])

        stuck_integral_column = IntegralColumn('StuckIntegralColumn',
                                               voltage_error_noise=8.0)
        stuck_integral_loop_writer = control_loop.ControlLoopWriter(
            'StuckIntegralColumn', [stuck_integral_column],
            namespaces=namespaces)
        stuck_integral_loop_writer.Write(argv[5], argv[6])


if __name__ == '__main__':
    argv = FLAGS(sys.argv)
    glog.init()
    sys.exit(main(argv))