# dim = 7
dim = int(random.random() * Dims)
data_mean = np.mean(measurements[dim, :])
# init 100 particles
N = 50
P = []
Estimation = []
for i in range(N):
x = particles()
x.set(random.gauss(data_mean, 0.2), 0.005, 5.0, 1)
P.append(x)
# gradient
g_1 = gradient_1(measurements[dim, :])
g_2 = gradient_2(measurements[dim, :])
g_1_bk = gradient_1_bk(measurements[dim, :])
g_2_bk = gradient_2_bk(measurements[dim, :])
# run motion model
for t in range(Length):
measurement = measurements[dim, t]
for n in range(N):
# P[n].motion(motionFCN, v=g_1[t], a=g_2[t], t=1)
P[n].motion(motionFCN, v=g_1_bk[t], a=0, t=1)
# P[n].motion(motionFCN, v=g_1_bk[t], a=g_2_bk[t], t=1)
pred_range = [64, 74]
# init 100 particles
N = 100
P = []
Estimation = []
for i in range(N):
x = particles()
x.set(random.gauss(data_mean, 1.0), 0.005, 5.0, 1)
P.append(x)
g_1_bk = gradient_1_bk(data[dim, :])
g_1_bk = average_gradient(g_1_bk, 3)
g_1 = gradient_1(data[dim, :])
g_2 = gradient_2(data[dim, :])
g_2_bk = gradient_2_bk(data[dim, :])
g_1_random = [random.gauss(0.0, 0.1) * .5 for x in range(Length)]
print(np.min(g_1_random), np.max(g_1_random))
print(np.min(g_1_bk), np.max(g_1_bk))
# motion with both velocity and acceleration
P1 = copy.deepcopy(P)
# prog = propagation(P1, motionFCN, data[dim, :], g_1_bk, g_2_bk)
prog = propagation(P1, motionFCN, data[dim, :], g_1_bk, None)
for t_idx in range(len(data[dim, :])):
prog.pred(t_idx)
from config import data_config, covariance_matrix
from matplotlib import pyplot as plt
# prepare data & choose arbitary dimension
data_path = os.path.join(data_config['root_dir'], data_config['data_name'])
data = normalize(readMat(data_path))
Dims, Length = data.shape
# Convert acceleration to covariance and weighted
g_1 = []
g_2 = []
# init gradient for all data, all channel
for k in range(8):
g_1.append(gradient_1(data[k, :]))
g_2.append(gradient_2(data[k, :]))
# convert list to array
g_1 = np.asarray(g_1)
g_2 = np.asarray(g_2)
g_2_update = copy.deepcopy(g_2)
# update gradient
for k in range(Length):
g_2_update[:, k] = update_acceleration(g_2[:, k], covariance_matrix)
print("FInish updateing acceleration")
print("Total Dimension is {}".format(Dims))