""" Created on Tue Jan 21 17:19:31 2020 This file focuses on the performance under different modes @author: lyh """ # Setting up the packages import numpy as np from matplotlib import pyplot as plt import ocs_io as tpio from performance import performance_tool # Setting up paths cfg_file = r"D:\research\TU Delft\programs\ALTCUBTOOL\cfg\parameters.cfg" cfg = tpio.ConfigFile(cfg_file) concept = 'Comb Constellation' #"Specular Constellation"#cfg.sim.concept (P_r1, P_a1, ac1, raw_data_rate1, proc_rate1) = performance_tool(cfg_file, concept, SAR_mode_enable=False, interleaved_mode_enable=False) (P_r2, P_a2, ac2, raw_data_rate2, proc_rate2) = performance_tool(cfg_file, concept, SAR_mode_enable=True, interleaved_mode_enable=False) (P_r3, P_a3, ac3, raw_data_rate3, proc_rate3) = performance_tool(cfg_file, concept,
def performance_tool(cfg_file, concept, SAR_mode_enable, interleaved_mode_enable): # INITIAL PARAMETERS cfg = tpio.ConfigFile(cfg_file) # lambda [m] l0 = const.c / cfg.radar.f0 # along tr. bw [rad] a_bw = l0 / cfg.radar.az_ant # cross tr. bw [rad] c_bw = l0 / cfg.radar.c_ant # along tr. eff. or smallest bw [deg] eff_a_bw = a_bw / np.sqrt(2) # cross tr. eff. or smallest bw [deg] eff_c_bw = c_bw / np.sqrt(2) # antenna gain gain_t = 10 * np.log10(4 * np.pi / a_bw / c_bw / 2) gain_r = gain_t # compressed pulse length pw = const.c / 2 / cfg.radar.bandwidth # calculate orbit alt0 = cfg.orbit.alt inc = cfg.orbit.inc e = cfg.orbit.e re_cycle_num = cfg.orbit.re_cycle_num re_day = cfg.orbit.re_day if np.size(alt0) > 1: alt = np.zeros_like(alt0) v = np.zeros_like(alt0) dis = np.zeros_like(alt0) for i in range(np.size(alt0)): alt[i], v[i], dis[i] = orbit_com(alt0[i], inc, e, re_cycle_num, re_day, cfg.orbit.con_repeat) else: alt, v, dis = orbit_com(alt0, inc, e, re_cycle_num, re_day) # mode recognize if concept == "Specular Constellation": # along tr. bw [rad] a_bw = l0 / cfg.radar.microsat_antenna_along # cross tr. bw [rad] c_bw = l0 / cfg.radar.microsat_antenna eff_a_bw = a_bw / np.sqrt(2) alt_m = alt / np.cos(a_bw / 2) gain_t = 10 * np.log10(4 * np.pi / a_bw / c_bw / 2) - 3 if cfg.sim.max_coverage: gain_r = 10 * np.log10(4 * np.pi / c_bw / c_bw / 2) if cfg.sim.interferometer_enable: # based on Ulaby model if consider the ocean surface is mirror alph=1 # at far range the SNR should be enough D = (np.tan(np.radians(cfg.radar.off_nadir)) * alt + c_bw / 2 * alt) * 2 off_nadir = np.arctan( (D / 2 + np.radians(cfg.radar.off_nadir) / 2 * alt) / alt) # maximal off-nadir angle gain_r = gain_r - 10 * np.log10(4.34 * np.degrees(off_nadir)) - 3 else: gain_r = gain_r - 3 if SAR_mode_enable: pulse_coh_bur = cfg.radar.pulse_coh_bur else: pulse_coh_bur = 1 if cfg.sim.plot_spec_coverage: if np.size(cfg.radar.microsat_antenna) > 1: ang = l0 / cfg.radar.microsat_antenna plt.figure() for i in range(np.size(alt_m)): plt.plot(cfg.radar.microsat_antenna, ang * cfg.orbit.alt[i] / 1000) leg = np.append( leg, ["Orbit height=%d km" % (cfg.orbit.alt[i] / 1000)]) plt.xticks(fontsize=16) plt.yticks(fontsize=16) plt.xlabel('Cross-track size of microcube (m)', fontsize=16) plt.ylabel('Coverage area (km)', fontsize=16) plt.title(concept) plt.legend(leg) # plt.ylim(2, 5) plt.grid() # calculation performance # Doppler BW [Hz] and prf Dop_bw = 2 * v / (l0 / eff_a_bw) if SAR_mode_enable: prf = 18700 else: prf = 4000 #int(1.4 * Dop_bw) prf_r = 1.4 * Dop_bw if cfg.sim.plot_prf: #plot the prf rule fig = plt.figure() ax1 = fig.add_subplot(111) ax1.plot(l0 / a_bw, prf_r) #ax1.set_xticks(fontsize=16) #ax1.set_yticks(fontsize=16) ax1.set_xlabel('Along-track antenna dimension of the Microsat (m)', fontsize=16) ax1.set_ylabel('PRF (Hz)', fontsize=16, color="blue") ax1.set_title("PRF vs. swath") ax2 = ax1.twinx() ax2.plot(l0 / a_bw, l0 / (1 / cfg.radar.microsat_antenna_along) * alt_m / 1000, 'r') ax2.set_ylabel('Cross-track swath length (km)', fontsize=16, color="red") plt.show() plt.grid() prt = 1 / prf # free space between pulses [m] #F_length = (prt - cfg.radar.pw_uc / 1000) / 2 * const.c # max # pulses in burst [ ] #max_pulse_burst = (alt * 2 / const.c / prt).astype(int) # along tr. footprint [km] al_res = eff_a_bw * alt # cross tr. footprint [km] # pulse interval time if SAR_mode_enable: bur_rep_int = cfg.radar.bur_rep_int else: bur_rep_int = prt * 1000 c_res = eff_c_bw * alt # distance between bursts [m] dis_burst = bur_rep_int * v / 1000 if cfg.sim.pulse_limited_enable: # pulse limited radius on ground [m] pul_rad = np.sqrt(2 * alt * pw) # Pulse lim. resolution [m] plu_res = 2 * pul_rad if SAR_mode_enable: # Synthetic Aperture length [m] syn_length = 1 / prf * pulse_coh_bur * v # SAR resolution [m] sar_res = prf * alt * l0 / (2 * pulse_coh_bur * v) # Raney limit [m] raney_l = 3 * pw * alt / sar_res # cells in Raney region cel = cell_calculate(al_res, raney_l, sar_res) # burst length [s] bur_length = pulse_coh_bur * prt # incoh. observ. per second N_incoh = cel / bur_rep_int * 1000 # incoh. observ. per res. cell N_incoh_r = cel * sar_res / v / bur_rep_int * 1000 else: # cells in Raney region cel = 1 sar_res = plu_res # burst length [s] bur_length = cfg.radar.pw_uc if np.size(dis_burst) > 1 and np.size(sar_res) > 1: N_incoh = np.zeros_like(dis_burst) for i in range(np.size(N_incoh)): if dis_burst[i] > 0.305 * alt[i] * l0 / sar_res[i]: N_incoh[i] = cel / bur_rep_int * 1000 else: N_incoh[i] = v / (0.305 * alt[i] * l0 / sar_res[i]) elif np.size(dis_burst) > 1 and np.size(sar_res) == 1: N_incoh = np.zeros_like(dis_burst) for i in range(np.size(N_incoh)): if dis_burst[i] > 0.305 * alt * l0 / sar_res: N_incoh[i] = cel / bur_rep_int[i] * 1000 else: N_incoh[i] = v / (0.305 * alt * l0 / sar_res) elif np.size(sar_res) > 1: N_incoh = np.zeros_like(sar_res) for i in range(np.size(sar_res)): if dis_burst > 0.305 * alt * l0 / sar_res[i]: N_incoh[i] = cel / bur_rep_int * 1000 else: N_incoh[i] = v / (0.305 * alt[i] * l0 / sar_res) else: if dis_burst > 0.305 * alt * l0 / sar_res: N_incoh = cel / bur_rep_int * 1000 else: N_incoh = v / (0.305 * alt * l0 / sar_res) N_incoh_r = N_incoh * sar_res / v # power performance KTF = -204 + cfg.loss.NF # illuminated area [m2] A = area_cal(sar_res, plu_res, pul_rad) A0 = area_cal(plu_res, plu_res, pul_rad) # Power calculation & accuracy sigma = 10 * np.log10(A) + cfg.loss.sigma0 sigma_w = 10 * np.log10(A0) + cfg.loss.sigma0 # req. peak power [W] if (np.size(cfg.loss.SNR) > 1): P_r = ((4 * np.pi)**3 * alt**4 * 10**((cfg.loss.SNR.reshape( (cfg.loss.SNR.shape + (1, ))) + cfg.loss.loss_in_at + KTF - gain_t - gain_r - sigma) / 10) / l0**2 / (cfg.radar.pw_uc / 1000) / pulse_coh_bur) # req. av. power [W] P_a = P_r * pulse_coh_bur * cfg.radar.pw_uc / bur_rep_int else: P_r = ((4 * np.pi)**3 * alt**4 * 10**((cfg.loss.SNR + cfg.loss.loss_in_at + KTF - gain_t - gain_r - sigma) / 10) / l0**2 / (cfg.radar.pw_uc / 1000) / pulse_coh_bur) # req. av. power [W] P_a = P_r * pulse_coh_bur * cfg.radar.pw_uc / bur_rep_int # calculating the SNR before the compression processing P_rc = cfg.radar.input_power * 10**( (gain_t + gain_r - cfg.loss.loss_in_at + sigma_w) / 10) * l0**2 / (4 * np.pi)**3 / alt**4 noise = 10**(KTF / 10) * cfg.radar.bandwidth SNR0 = 10 * np.log10(P_rc / noise) P_rcw = 10 * np.log10(P_rc) + 30 noise_w = 10 * np.log10(noise) + 30 # single cell height accuracy [cm] if (np.size(cfg.loss.SNR) > 1): ac = pw / np.sqrt(2 * N_incoh_r * (1 / (1 + 1 / 10**(cfg.loss.SNR.reshape( (cfg.loss.SNR.shape + (1, ))) / 10)))) * 100 else: ac = pw / np.sqrt(2 * N_incoh_r * (1 / (1 + 1 / 10**(cfg.loss.SNR / 10)))) * 100 # raw data rate [Mb/s], 16 bit samples raw_data_rate = pulse_coh_bur * cfg.radar.rg_num * 1000 / bur_rep_int / 1e6 * 16 # proc.per channel [kb/s] if SAR_mode_enable: proc_rate = v / sar_res * cfg.radar.rg_num * 16 / 1000 else: proc_rate = 4 * cfg.radar.rg_num * 16 / 1000 if cfg.sim.interferometer_enable: # cross track resolution [m] if np.size(cfg.loss.SNR) > 1: cross_res = 1 / np.sqrt( 10**(np.cos(a_bw / 2)**3 * cfg.loss.SNR.reshape( (cfg.loss.SNR.shape + (1, ))) / 10)) * l0 * alt / (2 * np.pi * cfg.radar.baseline) # This calculation is meaningless # height_ac = cross_res / np.cos(a_bw / 2) if interleaved_mode_enable: P_a = P_r * prf * cfg.radar.pw_uc / 1000 if SAR_mode_enable: N_incoh = cel / (pulse_coh_bur / prf) else: # cells in Raney region if 1 / prf * v > 0.305 * alt * l0 / sar_res: N_incoh = cel / pulse_coh_bur / prf else: N_incoh = min(v / (0.305 * alt * l0 / sar_res), prf) # incoh. observ. per res. cell N_incoh_r = cel * sar_res / v / pulse_coh_bur * prf ac = pw / np.sqrt(2 * N_incoh_r * (1 / (1 + 1 / 10**(cfg.loss.SNR.reshape( (cfg.loss.SNR.shape + (1, ))) / 10)))) * 100 raw_data_rate = prf * cfg.radar.rg_num / 1e6 * 16 # analysis of cells in the raney region if cfg.sim.plot_raney: plt.figure() plt.plot(l0 / a_bw, 2 * raney_l / 1000) plt.xlabel('Along-track antenna dimension of the Microsat (m)', fontsize=16) plt.ylabel('Along-track footprint (km)', fontsize=16) plt.plot(l0 / a_bw, al_res / 1000, 'r') ax2.set_ylabel('Length (km)', fontsize=16) plt.legend(['Raney region length', 'Along-track footprint']) plt.show() plt.grid() fig = plt.figure() ax1 = fig.add_subplot(111) ax1.plot(l0 / a_bw, cel) #ax1.set_xticks(fontsize=16) #ax1.set_yticks(fontsize=16) ax1.set_xlabel('Along-track antenna dimension of the Microsat (m)', fontsize=16) ax1.set_ylabel('cell number in the raney region', fontsize=16, color="blue") ax2 = ax1.twinx() ax2.plot(l0 / a_bw, ac, 'r') ax2.set_ylabel('SSH accuracy', fontsize=16, color="red") plt.show() plt.grid() return P_r, P_a, ac, raw_data_rate, proc_rate