# ERR INIT switch_r_nr
print ('\n ERR INIT switch_r_nr \nswitch_r_nr must be 1 or 0 !!!')
# </editor-fold>
# test
plt.plot(sampled_data_FC1[:,1])
plt.plot(sampled_data_FC2[:,1])
plt.show()
## use the svr fit the FC1-FC2 and get the FC2 predict value
FC2_prog_pred=svr_p.svr_predict(sampled_data_FC1,sampled_data_FC2,len_real=len_prognosis_FC2,len_train=len_learn_FC2)
## get the init parameter
para_FC1=sip.get_init_para(sampled_data_FC1,re_calc_len=len_prognosis_FC1)
para_FC2=sip.get_init_para(sampled_data_FC2,re_calc_len=len_prognosis_FC2)
#para_FC1_minus_FC2=sip.get_init_para(sampled_data_FC1_minus_FC2,re_calc_len=len_prognosis_FC2)
print ('\n get the init parameter successfully!')
## prognostics method 1: particle filter to parameter identification
# -----------------------------------------------------------------------------------
### model_linear good parameter, should save
[[filter_data,real_fil],[predictions_data,real_pred],[fil_pred,real_all]]=rp_ps.Run_ParticleFilter(
sampled_data=sampled_data_FC2,
x_svr_pred=FC2_prog_pred,para=para_FC2,
x_process_noise_sigma=0.2,
x is the system's state, and it's derivatives are some settings about the state.
alpha, beta are the parameters, and their derivations are some settings about the parameters.
z is the system's observation.
'''
import numpy as np
import pf_class as PF_C
import matplotlib.pyplot as plt
import read_sam_data as rsd
import solve_init_para as sip
len_prognostic = 1000
sampled_data_FC1=rsd.re_sam_data(FC='FC1')
sampled_data_FC2=rsd.re_sam_data(FC='FC2')
para=sip.get_init_para(sampled_data_FC1,re_calc_len=len_prognostic)
def Run_ParticleFilter(sampled_data,para,
x_process_noise_sigma=0.02,beta_process_noise_sigma=0.02,
sensor_std_err = 0.01,
x_init_gauss_sigma=0.1,beta_init_gauss_sigma=0.1,
x_init_uniform_bound=0.1,beta_init_uniform_bound=0.01,
len_prognostic=len_prognostic,num_particles=2000,iterations=100,
):
t_real = sampled_data[0:len_prognostic,0]
x_obs= sampled_data[0:len_prognostic,1]
para_0=para[0][4]
para_linear_beta=para[1][4,:]
x_init_range=np.array([x_obs[0]-x_init_uniform_bound,x_obs[0]+x_init_uniform_bound])
beta_init_range=np.array([para_0-beta_init_uniform_bound,para_0+beta_init_uniform_bound])