def Obj(x):
Obj = 0
for k in range(len(Stancet[:, 0])):
L0 = Stancex[k, 0]
L1 = Stancex[k, -1]
T = Stancet[k, -1]
Fit = polynomial_value(L0, L1, T, Stancet[k, :], x)
Diff_Swingx = Stancex[k, :] - Fit
Disp_end = (Stancex[k, -1] - Fit[-1])**2
Velo_start = ((Stancex[k, 1]-Stancex[k, 0])/(Stancet[k, 1]-Stancet[k, 0])
- (Fit[1]-Fit[0])/(Stancet[k, 1]-Stancet[k, 0]))**2
Velo_end = ((Stancex[k, -1]-Stancex[k, -2])/(Stancet[k, -1]-Stancet[k, -2])
- (Fit[-1]-Fit[-2])/(Stancet[k, -1]-Stancet[k, -2]))**2
Obj = Obj + np.sum(Diff_Swingx**2) + (500*Velo_start + 1000*Disp_end
+ 500*Velo_end)
Obj = Obj/len(Stancet[:, 0])
return Obj
def get_stancex_polynomial(Lstancex, Rstancex, Lstancet, Rstancet, num_par,
repeat, store_path):
LEd = len(Lstancex[:, 0])
REd = len(Rstancex[:, 0])
fig = plt.figure(figsize=(6, 6))
for k in range(LEd):
plt.plot(Lstancet[k, :], Lstancex[k, :], 'r-')
for j in range(REd):
plt.plot(Rstancet[j, :], Rstancex[j, :], 'b-')
fig.savefig(store_path + '/Path_Stancex_plot.png')
Ed = LEd + REd
global Stancex
global Stancet
Stancex = np.vstack((Lstancex, Rstancex))
Stancet = np.vstack((Lstancet, Rstancet))
Result = np.zeros((repeat, num_par))
for m in range(repeat):
np.random.seed()
x0 = 10 - 20*np.random.random(num_par)
res = minimize(Obj, x0, method='BFGS', options={'disp': True})
Result[m, :] = res['x']
Coef = np.average(Result, axis=0)
Coef_st = np.std(Result, axis=0)
RMS = np.zeros(Ed)
for k in range(Ed):
L0 = Stancex[k, 0]
L1 = Stancex[k, -1]
T = Stancet[k, -1]
Fit = polynomial_value(L0, L1, T, Stancet[k, :], Coef)
RMS[k] = np.mean(np.square(Stancex[k, :] - Fit))
with open(store_path + '/RMS_Stancex_normalized'+str(num_par)+'.txt', 'w') as outfile:
StringP = ""
for i in range(len(RMS)):
StringP += str(RMS[i])
StringP += "\n"
outfile.write(StringP)
with open(store_path + '/Coe_Stancex_normalized'+str(num_par)+'.txt', 'w') as outfile:
StringP = ""
for i in range(len(Result[:, 0])):
for j in range(len(Result[0, :])):
StringP += str(Result[i, j])
StringP += " "
StringP += "\n"
outfile.write(StringP)
with open(store_path + '/Coe_Stancex_normalized'+str(num_par)+'_std.txt', 'w') as outfile:
StringP = ""
for j in range(len(Result[0, :])):
StringP += str(Coef[j])
StringP += " "
StringP += str(Coef_st[j])
StringP += "\n"
outfile.write(StringP)
index_max = RMS.argmax()
index_min = RMS.argmin()
index_med = np.argsort(RMS)[len(RMS)//2]
L0 = Stancex[index_min, 0]
L1 = Stancex[index_min, -1]
T = Stancet[index_min, -1]
L01 = Stancex[index_med, 0]
L11 = Stancex[index_med, -1]
T1 = Stancet[index_med, -1]
L02 = Stancex[index_max, 0]
L12 = Stancex[index_max, -1]
T2 = Stancet[index_max, -1]
Fit = polynomial_value(L0, L1, T, Stancet[index_min, :], Coef)
Fit1 = polynomial_value(L01, L11, T1, Stancet[index_med, :], Coef)
Fit2 = polynomial_value(L02, L12, T2, Stancet[index_max, :], Coef)
fig1 = plt.figure(figsize=(12, 8))
matplotlib.rc('xtick', labelsize=16)
matplotlib.rc('ytick', labelsize=16)
ax1 = fig1.add_subplot(1,3,1)
plt.xlabel('t (s)', fontsize=16)
plt.ylabel('Disp (m)', fontsize=16)
plt.title('Best Fit', fontsize=16)
ax1.plot(Stancet[index_min, :], Stancex[index_min, :], 'r-')
ax1.plot(Stancet[index_min, :], Fit, 'g.')
ax2 = fig1.add_subplot(1,3,2)
plt.xlabel('t (s)', fontsize=16)
plt.title('Median Fit', fontsize=16)
ax2.plot(Stancet[index_med, :], Stancex[index_med, :], 'r-')
ax2.plot(Stancet[index_med, :], Fit1, 'g.')
ax3 = fig1.add_subplot(1,3,3)
plt.xlabel('t (s)', fontsize=16)
plt.title('Worst Fit', fontsize=16)
ax3.plot(Stancet[index_max, :], Stancex[index_max, :], 'r-', label = 'experiment data')
ax3.plot(Stancet[index_max, :], Fit2, 'g.', label = 'polynomial fit')
plt.legend(fontsize=13)
fig1.savefig(store_path + '/Fit_Stancex_info.png')