SemiUnitaryDifferentialMIMO-IdealRayleigh-BER.py¶

This webpage introduces an API example for the noncohrent scenario, which uses SemiUnitaryDifferentialMLDSimulator. Other examples are found in SemiUnitaryDifferentialMLDSimulatorTest.

This example compares the BER performance of the differential star-QAM (SQAM), differential orthogonal space-time block code (DOSTBC) with SQAM, and diagonal unitary code (DUC) schemes. The simulation parameters are given in the table below.

Performance Results¶

../_images/SemiUnitaryDifferentialMIMO-IdealRayleigh-BER.svg

Simulation Parameters¶

Parameter	Value
Channel	Ideal Rayleigh fading
Number of transmit antennas	\(M=2\)
Number of receive antennas	\(N=2\)
Constellation size	\(L=16,256\)
Transmission rate	\(R=B/T=8/2=4\) [bit/symbol]

Reproducible Code¶

import sys
import numpy as np
import matplotlib.pyplot as plt
from imtoolkit import Parameters, Modulator, OSTBCode, DiagonalUnitaryCode, TASTCode, IdealRayleighChannel, DifferentialMLDSimulator, SemiUnitaryDifferentialMLDSimulator

plt.switch_backend('agg')
plt.rcParams['xtick.direction'] = 'in'
plt.rcParams['ytick.direction'] = 'in'
plt.rcParams['markers.fillstyle'] = 'none'

def simulateBER(argstr):
    params = Parameters(argstr)
    if params.code == "symbol":
        codes = np.array(Modulator(params.mod, params.L).symbols).reshape(params.L, 1, 1)
    elif params.code == "OSTBC":
        codes = OSTBCode(params.M, params.mod, params.L).codes
    elif params.code == "DUC":
        codes = DiagonalUnitaryCode(params.M, params.L).codes

    channel = IdealRayleighChannel(params.ITi, params.M, params.N)

    if params.sim == "diff":
        sim = DifferentialMLDSimulator(codes, channel)
    elif params.sim == "sudiff":
        sim = SemiUnitaryDifferentialMLDSimulator(codes, channel)

    return sim.simulateBERParallel(params, outputFile=False, printValue=False)

if __name__ == '__main__':
    fig, ax = plt.subplots()
    ax.set_xlabel("SNR [dB]")
    ax.set_ylabel("BER")
    ax.set_xlim(0, 40)
    plt.ylim(1e-7, 1e0)
    plt.yscale("log")
    ax.tick_params(pad=8)

    ret = simulateBER("BERP_sim=sudiff_channel=rayleigh_code=symbol_M=1_N=2_L=16_mod=SQAM_ITo=1e3_ITi=1e5_snrfrom=0.00_to=40.00_len=21")
    ax.plot(ret["snr_dB"], ret["ber"], color="k", marker="s", linestyle="-", label="Differential SQAM [1]")

    ret = simulateBER("BERP_sim=sudiff_channel=rayleigh_code=OSTBC_M=2_N=2_T=2_L=16_mod=SQAM_ITo=1e3_ITi=1e5_snrfrom=0.00_to=40.00_len=21")
    ax.plot(ret["snr_dB"], ret["ber"], color="b", marker="o", linestyle="-", label="DOSTBC w/ SQAM [2]")

    ret = simulateBER("BERP_sim=diff_channel=rayleigh_code=DUC_M=2_N=2_T=2_L=256_ITo=1e3_ITi=1e5_snrfrom=0.00_to=40.00_len=21")
    ax.plot(ret["snr_dB"], ret["ber"], color="r", marker="^", linestyle="-", label="DUC [3]")

    handles, labels = ax.get_legend_handles_labels()
    legend = ax.legend(handles, labels, loc="best", frameon=True)
    frame = legend.get_frame()
    frame.set_facecolor('white')
    frame.set_edgecolor('white')

    #plt.show()
    plt.savefig(sys.argv[0].replace(".py", ".svg"))

Fork me on GitHub