blob: 175fffab490144e96cc9c5908b378ad56d9d2f63 [file] [log] [blame]
#!/usr/bin/python
from aos.util.trapezoid_profile import TrapezoidProfile
from frc971.control_loops.python import control_loop
from frc971.control_loops.python import controls
import numpy
from matplotlib import pylab
import glog
class LinearSystemParams(object):
def __init__(self,
name,
motor,
G,
radius,
mass,
q_pos,
q_vel,
kalman_q_pos,
kalman_q_vel,
kalman_q_voltage,
kalman_r_position,
dt=0.005):
self.name = name
self.motor = motor
self.G = G
self.radius = radius
self.mass = mass
self.q_pos = q_pos
self.q_vel = q_vel
self.kalman_q_pos = kalman_q_pos
self.kalman_q_vel = kalman_q_vel
self.kalman_q_voltage = kalman_q_voltage
self.kalman_r_position = kalman_r_position
self.dt = dt
class LinearSystem(control_loop.ControlLoop):
def __init__(self, params, name='LinearSystem'):
super(LinearSystem, self).__init__(name)
self.params = params
self.motor = params.motor
# Gear ratio
self.G = params.G
self.radius = params.radius
# Mass in kg
self.mass = params.mass + self.motor.motor_inertia / (
(self.G * self.radius)**2.0)
# Control loop time step
self.dt = params.dt
# State is [position, velocity]
# Input is [Voltage]
C1 = self.motor.Kt / (self.G * self.G * self.radius * self.radius *
self.motor.resistance * self.mass * self.motor.Kv)
C2 = self.motor.Kt / (
self.G * self.radius * self.motor.resistance * self.mass)
self.A_continuous = numpy.matrix([[0, 1], [0, -C1]])
# Start with the unmodified input
self.B_continuous = numpy.matrix([[0], [C2]])
glog.debug(repr(self.A_continuous))
glog.debug(repr(self.B_continuous))
self.C = numpy.matrix([[1, 0]])
self.D = numpy.matrix([[0]])
self.A, self.B = self.ContinuousToDiscrete(self.A_continuous,
self.B_continuous, self.dt)
controllability = controls.ctrb(self.A, self.B)
glog.debug('Controllability of %d',
numpy.linalg.matrix_rank(controllability))
glog.debug('Mass: %f', self.mass)
glog.debug('Stall force: %f',
self.motor.stall_torque / self.G / self.radius)
glog.debug('Stall acceleration: %f',
self.motor.stall_torque / self.G / self.radius / self.mass)
glog.debug('Free speed is %f',
-self.B_continuous[1, 0] / self.A_continuous[1, 1] * 12.0)
self.Q = numpy.matrix([[(1.0 / (self.params.q_pos**2.0)), 0.0],
[0.0, (1.0 / (self.params.q_vel**2.0))]])
self.R = numpy.matrix([[(1.0 / (12.0**2.0))]])
self.K = controls.dlqr(self.A, self.B, self.Q, self.R)
q_pos_ff = 0.005
q_vel_ff = 1.0
self.Qff = numpy.matrix([[(1.0 / (q_pos_ff**2.0)), 0.0],
[0.0, (1.0 / (q_vel_ff**2.0))]])
self.Kff = controls.TwoStateFeedForwards(self.B, self.Qff)
glog.debug('K %s', repr(self.K))
glog.debug('Poles are %s',
repr(numpy.linalg.eig(self.A - self.B * self.K)[0]))
self.Q = numpy.matrix([[(self.params.kalman_q_pos**2.0), 0.0],
[0.0, (self.params.kalman_q_vel**2.0)]])
self.R = numpy.matrix([[(self.params.kalman_r_position**2.0)]])
self.KalmanGain, self.Q_steady = controls.kalman(
A=self.A, B=self.B, C=self.C, Q=self.Q, R=self.R)
glog.debug('Kal %s', repr(self.KalmanGain))
# The box formed by U_min and U_max must encompass all possible values,
# or else Austin's code gets angry.
self.U_max = numpy.matrix([[12.0]])
self.U_min = numpy.matrix([[-12.0]])
self.InitializeState()
class IntegralLinearSystem(LinearSystem):
def __init__(self, params, name='IntegralLinearSystem'):
super(IntegralLinearSystem, self).__init__(params, name=name)
self.A_continuous_unaugmented = self.A_continuous
self.B_continuous_unaugmented = self.B_continuous
self.A_continuous = numpy.matrix(numpy.zeros((3, 3)))
self.A_continuous[0:2, 0:2] = self.A_continuous_unaugmented
self.A_continuous[0:2, 2] = self.B_continuous_unaugmented
self.B_continuous = numpy.matrix(numpy.zeros((3, 1)))
self.B_continuous[0:2, 0] = self.B_continuous_unaugmented
self.C_unaugmented = self.C
self.C = numpy.matrix(numpy.zeros((1, 3)))
self.C[0:1, 0:2] = self.C_unaugmented
self.A, self.B = self.ContinuousToDiscrete(self.A_continuous,
self.B_continuous, self.dt)
self.Q = numpy.matrix([[(self.params.kalman_q_pos**2.0), 0.0, 0.0],
[0.0, (self.params.kalman_q_vel**2.0), 0.0],
[0.0, 0.0, (self.params.kalman_q_voltage**2.0)]])
self.R = numpy.matrix([[(self.params.kalman_r_position**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.K_unaugmented = self.K
self.K = numpy.matrix(numpy.zeros((1, 3)))
self.K[0, 0:2] = self.K_unaugmented
self.K[0, 2] = 1
self.Kff = numpy.concatenate(
(self.Kff, numpy.matrix(numpy.zeros((1, 1)))), axis=1)
self.InitializeState()
def RunTest(plant,
end_goal,
controller,
observer=None,
duration=1.0,
use_profile=True,
kick_time=0.5,
kick_magnitude=0.0,
max_velocity=0.3,
max_acceleration=10.0):
"""Runs the plant with an initial condition and goal.
Args:
plant: plant object to use.
end_goal: end_goal state.
controller: LinearSystem object to get K from, or None if we should
use plant.
observer: LinearSystem object to use for the observer, or None if we
should use the actual state.
duration: float, time in seconds to run the simulation for.
kick_time: float, time in seconds to kick the robot.
kick_magnitude: float, disturbance in volts to apply.
max_velocity: float, the max speed in m/s to profile.
max_acceleration: float, the max acceleration in m/s/s to profile.
"""
t_plot = []
x_plot = []
v_plot = []
a_plot = []
x_goal_plot = []
v_goal_plot = []
x_hat_plot = []
u_plot = []
offset_plot = []
if controller is None:
controller = plant
vbat = 12.0
goal = numpy.concatenate((plant.X, numpy.matrix(numpy.zeros((1, 1)))),
axis=0)
profile = TrapezoidProfile(plant.dt)
profile.set_maximum_acceleration(max_acceleration)
profile.set_maximum_velocity(max_velocity)
profile.SetGoal(goal[0, 0])
U_last = numpy.matrix(numpy.zeros((1, 1)))
iterations = int(duration / plant.dt)
for i in xrange(iterations):
t = i * plant.dt
observer.Y = plant.Y
observer.CorrectObserver(U_last)
offset_plot.append(observer.X_hat[2, 0])
x_hat_plot.append(observer.X_hat[0, 0])
next_goal = numpy.concatenate(
(profile.Update(end_goal[0, 0], end_goal[1, 0]),
numpy.matrix(numpy.zeros((1, 1)))),
axis=0)
ff_U = controller.Kff * (next_goal - observer.A * goal)
if use_profile:
U_uncapped = controller.K * (goal - observer.X_hat) + ff_U
x_goal_plot.append(goal[0, 0])
v_goal_plot.append(goal[1, 0])
else:
U_uncapped = controller.K * (end_goal - observer.X_hat)
x_goal_plot.append(end_goal[0, 0])
v_goal_plot.append(end_goal[1, 0])
U = U_uncapped.copy()
U[0, 0] = numpy.clip(U[0, 0], -vbat, vbat)
x_plot.append(plant.X[0, 0])
if v_plot:
last_v = v_plot[-1]
else:
last_v = 0
v_plot.append(plant.X[1, 0])
a_plot.append((v_plot[-1] - last_v) / plant.dt)
u_offset = 0.0
if t >= kick_time:
u_offset = kick_magnitude
plant.Update(U + u_offset)
observer.PredictObserver(U)
t_plot.append(t)
u_plot.append(U[0, 0])
ff_U -= U_uncapped - U
goal = controller.A * goal + controller.B * ff_U
if U[0, 0] != U_uncapped[0, 0]:
profile.MoveCurrentState(numpy.matrix([[goal[0, 0]], [goal[1, 0]]]))
glog.debug('Time: %f', t_plot[-1])
glog.debug('goal_error %s', repr(end_goal - goal))
glog.debug('error %s', repr(observer.X_hat - end_goal))
pylab.subplot(3, 1, 1)
pylab.plot(t_plot, x_plot, label='x')
pylab.plot(t_plot, x_hat_plot, label='x_hat')
pylab.plot(t_plot, x_goal_plot, label='x_goal')
pylab.legend()
pylab.subplot(3, 1, 2)
pylab.plot(t_plot, u_plot, label='u')
pylab.plot(t_plot, offset_plot, label='voltage_offset')
pylab.legend()
pylab.subplot(3, 1, 3)
pylab.plot(t_plot, a_plot, label='a')
pylab.legend()
pylab.show()
def PlotStep(params, R):
"""Plots a step move to the goal.
Args:
R: numpy.matrix(2, 1), the goal"""
plant = LinearSystem(params, params.name)
controller = IntegralLinearSystem(params, params.name)
observer = IntegralLinearSystem(params, params.name)
# Test moving the system.
initial_X = numpy.matrix([[0.0], [0.0]])
augmented_R = numpy.matrix(numpy.zeros((3, 1)))
augmented_R[0:2, :] = R
RunTest(
plant,
end_goal=augmented_R,
controller=controller,
observer=observer,
duration=2.0,
use_profile=False,
kick_time=1.0,
kick_magnitude=0.0)
def PlotKick(params, R):
"""Plots a step motion with a kick at 1.0 seconds.
Args:
R: numpy.matrix(2, 1), the goal"""
plant = LinearSystem(params, params.name)
controller = IntegralLinearSystem(params, params.name)
observer = IntegralLinearSystem(params, params.name)
# Test moving the system.
initial_X = numpy.matrix([[0.0], [0.0]])
augmented_R = numpy.matrix(numpy.zeros((3, 1)))
augmented_R[0:2, :] = R
RunTest(
plant,
end_goal=augmented_R,
controller=controller,
observer=observer,
duration=2.0,
use_profile=False,
kick_time=1.0,
kick_magnitude=2.0)
def PlotMotion(params, R, max_velocity=0.3, max_acceleration=10.0):
"""Plots a trapezoidal motion.
Args:
R: numpy.matrix(2, 1), the goal,
max_velocity: float, The max velocity of the profile.
max_acceleration: float, The max acceleration of the profile.
"""
plant = LinearSystem(params, params.name)
controller = IntegralLinearSystem(params, params.name)
observer = IntegralLinearSystem(params, params.name)
# Test moving the system.
initial_X = numpy.matrix([[0.0], [0.0]])
augmented_R = numpy.matrix(numpy.zeros((3, 1)))
augmented_R[0:2, :] = R
RunTest(
plant,
end_goal=augmented_R,
controller=controller,
observer=observer,
duration=2.0,
use_profile=True,
max_velocity=max_velocity,
max_acceleration=max_acceleration)
def WriteLinearSystem(params, plant_files, controller_files, year_namespaces):
"""Writes out the constants for a linear system to a file.
Args:
params: LinearSystemParams, the parameters defining the system.
plant_files: list of strings, the cc and h files for the plant.
controller_files: list of strings, the cc and h files for the integral
controller.
year_namespaces: list of strings, the namespace list to use.
"""
# Write the generated constants out to a file.
linear_system = LinearSystem(params, params.name)
loop_writer = control_loop.ControlLoopWriter(
linear_system.name, [linear_system], namespaces=year_namespaces)
loop_writer.AddConstant(
control_loop.Constant('kFreeSpeed', '%f',
linear_system.motor.free_speed))
loop_writer.AddConstant(
control_loop.Constant('kOutputRatio', '%f',
linear_system.G * linear_system.radius))
loop_writer.AddConstant(
control_loop.Constant('kRadius', '%f', linear_system.radius))
loop_writer.Write(plant_files[0], plant_files[1])
integral_linear_system = IntegralLinearSystem(params,
'Integral' + params.name)
integral_loop_writer = control_loop.ControlLoopWriter(
integral_linear_system.name, [integral_linear_system],
namespaces=year_namespaces)
integral_loop_writer.Write(controller_files[0], controller_files[1])