double_pendulum/examples/ideal/pfl.py at main · based-robotics/double_pendulum · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
import sys
import os
from datetime import datetime
import yaml
import numpy as np

from double_pendulum.model.symbolic_plant import SymbolicDoublePendulum
from double_pendulum.model.model_parameters import model_parameters
from double_pendulum.simulation.simulation import Simulator
from double_pendulum.utils.plotting import plot_timeseries
from double_pendulum.utils.csv_trajectory import save_trajectory
from double_pendulum.controller.partial_feedback_linearization.symbolic_pfl import (SymbolicPFLController,
                                                                                    SymbolicPFLAndLQRController)

# model parameters
design = "design_A.0"
model = "model_2.0"
robot = "acrobot"

pfl_method = "collocated"
with_lqr = True

if robot == "acrobot":
    torque_limit = [0.0, 5.0]
    active_act = 1
if robot == "pendubot":
    torque_limit = [5.0, 0.0]
    active_act = 0

model_par_path = "../../data/system_identification/identified_parameters/"+design+"/"+model+"/model_parameters.yml"
mpar = model_parameters(filepath=model_par_path)
mpar.set_motor_inertia(0.)
mpar.set_damping([0., 0.])
mpar.set_cfric([0., 0.])
mpar.set_torque_limit(torque_limit)

# simulation parameters
integrator = "runge_kutta"
goal = [np.pi, 0., 0., 0.]
dt = 0.01
x0 = [0.1, 0.0, 0.0, 0.0]
t_final = 10.0

# controller parameters
if robot == "acrobot":
    # lqr parameters
    Q = np.diag((0.97, 0.93, 0.39, 0.26))
    R = np.diag((0.11, 0.11))
    if pfl_method == "collocated":
        #par = [6.78389278, 5.66430937, 9.98022384]
        #par = [1.58316202e-03, 2.94951787e+00, 1.44919303e+00]
        #par = [19.95373044, 14.76768604, 18.23010249]
        #par = [9.83825279, 9.42196979, 7.56036347]
        par = [0.0093613, 0.99787652, 0.9778557 ]  # new2
        #par = [4.95841985, 6.15434537, 9.54086796]  # new2
    elif pfl_method == "noncollocated":
        par = [9.19534629, 2.24529733, 5.90567362]
elif robot == "pendubot":
    # lqr parameters
    Q = np.diag([0.00125, 0.65, 0.000688, 0.000936])
    R = np.diag([25.0, 25.0])
    if pfl_method == "collocated":
        #par = [8.0722899, 4.92133648, 3.53211381]
        par = [8.8295605, 6.78718988, 4.42965278]
        #par = [5.23776024, 4.87113077, 3.0001595]
    elif pfl_method == "noncollocated":
        #par = [26.34039456, 99.99876263, 11.89097532]
        par = [8.0722899, 4.92133648, 3.53211381]

plant = SymbolicDoublePendulum(model_pars=mpar)

if with_lqr:
    controller = SymbolicPFLAndLQRController(model_pars=mpar,
                                             robot=robot,
                                             pfl_method=pfl_method)
    controller.lqr_controller.set_cost_parameters(p1p1_cost=Q[0, 0],
                                                  p2p2_cost=Q[1, 1],
                                                  v1v1_cost=Q[2, 2],
                                                  v2v2_cost=Q[3, 3],
                                                  p1v1_cost=0.,
                                                  p1v2_cost=0.,
                                                  p2v1_cost=0.,
                                                  p2v2_cost=0.,
                                                  u1u1_cost=R[0, 0],
                                                  u2u2_cost=R[1, 1],
                                                  u1u2_cost=0.)
    # controller.lqr_controller.set_parameters(failure_value=np.nan,
    #                                          cost_to_go_cut=1000)
else:  # without lqr
    controller = SymbolicPFLController(model_pars=mpar,
                                       robot=robot,
                                       pfl_method=pfl_method)

sim = Simulator(plant=plant)

controller.set_goal(goal)
controller.set_cost_parameters_(par)

print(f"Simulating {pfl_method} PFL controller for {robot}")
print(f"LQR: {with_lqr}")
print(f"dt: {dt}")
print(f"t final: {t_final}")
print(f"Cost parameters: {par}")

controller.init()

T, X, U = sim.simulate_and_animate(t0=0.0,
                                   x0=x0,
                                   tf=t_final,
                                   dt=dt,
                                   controller=controller,
                                   integrator=integrator,
                                   phase_plot=False,
                                   save_video=False)

energy = controller.en
des_energy = controller.desired_energy

# saving and plotting
timestamp = datetime.today().strftime("%Y%m%d-%H%M%S")
save_dir = os.path.join("data", design, model, robot, "pfl", pfl_method, timestamp)
os.makedirs(save_dir)

save_trajectory(csv_path=os.path.join(save_dir, "trajectory.csv"),
                T=T,
                X=X,
                U=U)
mpar.save_dict(os.path.join(save_dir, "model_parameters.yml"))
controller.save(save_dir)

plot_timeseries(T=T, X=X, U=U, energy=energy,
                plot_energy=True,
                pos_y_lines=[-np.pi, np.pi],
                tau_y_lines=[-torque_limit[active_act], torque_limit[active_act]],
                energy_y_lines=[des_energy],
                save_to=os.path.join(save_dir, "time_series"))