Proc. SPIE. 9619, 2015 International Conference on Optical Instruments and Technology: Optoelectronic Devices and Optical Signal Processing
KEYWORDS: Signal to noise ratio, Transmitters, Digital signal processing, Error analysis, Interference (communication), Receivers, Phase shift keying, Monte Carlo methods, Signal processing, Optical communications
In coherent optical phase-shift-keying (PSK) system, the conventional decision-aided maximum likelihood (DA-ML) phase estimator only considers constant phase noise over the entire observation interval. In this paper, we take into account the time-varying phase noise and propose a flexible DA-ML phase estimator for carrier phase estimation. The weighted coefficients are introduced to estimate carrier phase more accurately than the conventional scheme. The phase estimation error of this flexible DA-ML receiver is analyzed and compared with results from the conventional DA-ML receiver. The analytical bit error rate (BER) follows well with Monte Carlo (MC) simulations. Simulation results also show that by using the flexible DA-ML receiver in quadrature PSK (QPSK) system, bit signal-to-noise ratio (SNR) improvement over the conventional DA-ML receiver at BER of 10<sup>-3 </sup>is 0.2dB and 1.2dB with block length of 20 and 50, respectively, when the laser linewidth is 2MHz.
We study the limitation of electrical bandwidth in QPSK optical coherent receiver with decision-aided maximum likelihood (DAML) carrier phase estimation. Before signal sampling, a low-pass rectangular filter is utilized to model the bandwidth limitation of electronic devices in the receiver. On one hand the limited bandwidth can reduce the power of additive noise, on the other hand it would distort the signal since inter-symbol interference (ISI) between adjacent symbols occurs. In this paper we discuss the effects on signal amplitude and phase caused by ISI and find original distribution of QPSK constellation points will shift due to ISI. When adjacent symbol interference, that is, the ISI effect of one previous symbol, is taken into account, the distribution changes into the form of 16-QAM, while changing as 4K+1-QAM if K previous symbols are considered. The impact on the subsequent DAML algorithm and the final bit error rate (BER) calculation due to constellation points shifts are analyzed intensively. Monte-Carlo (MC) simulation results show that 0.7 to 0.9 times symbol rate is a reasonable bandwidth range for different optical signal to noise ratios (OSNRs), and optimum bandwidth is often inside this range. Results also reveal that filter-induced ISI would degrade the BER performance of DAML receiver when OSNR is over 10 dB.
Proc. SPIE. 9270, Optoelectronic Devices and Integration V
KEYWORDS: Signal to noise ratio, Visible radiation, Light emitting diodes, Modulation, Reliability, Receivers, Control systems, Telecommunications, Orthogonal frequency division multiplexing, Signal detection
In traditional dimming control system using pulse width modulation (PWM) combined with M-QAM OFDM scheme,
OFDM signal is only transmitted during “on” period. To guarantee the communication quality, reduction of duty cycle
will cause increased symbol rate or added LED power. This means system BER performance degradation and power
consumption. In order to solve the defects of the traditional dimming scheme, we propose a new dimming control
scheme in indoor visible light communication, which combines OFDM signal and multi-pulse position modulation
(MPPM) light pulse well with each other. By means of dividing traditional PWM pulses into MPPM pulses with the
same duty cycle, the pattern effect of MPPM pulses is utilized, which makes excess signal transmission possible.
Simulation results show that when reducing the brightness of LED the achievable symbol rate using dimming control
patterns is not higher than the traditional PWM scheme and the LED power is also reduced, which satisfies both system
reliability and energy effectiveness under constant high data rate and BER less than 10<sup>-3</sup>.