Estimation and detection of signals in a turbulent free space optical communications channel using array detectors

Free space optical (FSO) communications is the transmission of information via laser light, either through the atmosphere or through space. The detection of FSO communications signals relies on knowledge of the received optical signal intensity, which fluctuates as a result of transmission through the turbulent atmosphere. This intensity fluctuation can result in degraded signal quality and/or deep fades in which the signal is essentially lost. Methods can be employed to combat the effects of turbulence and provide reliable communications via FSO communications. In this research we investigate means of optimally detecting FSO communications where the signal intensity is a known quantity. Towards this end we suggest an optimal signal intensity estimator as well as receiver architectures which use intensity estimates in their implementations.; Estimators motivated by the minimum mean square error (MMSE) estimation criterion are presented for the estimation of optical signal intensity in FSO communication systems impaired by atmospheric-induced scintillation. Estimators are considered when the received signal is detected using either a p-i-n photodiode or an avalanche photodiode (APD). Further, both single element detectors and an array of photodetectors are considered for the detection of the optical signal. When considering the array of photodetectors, joint parameter estimators for the estimation of signal intensities across a photo-detector array are investigated. Three scenarios are considered for both the single element detector and for the array detector. First, it is assumed that the received signal is detected using a p-i-n photodiode or p-i-n photodiode array and that the receiver operates under a shot-noise-limited condition. In the second scenario, the received signal is detected using an APD or an APD array and the receiver operates under a shot-noise-limited condition. In the third scenario, the received signal is detected using an APD or an APD array, but the receiver is impaired by thermal noise as well as shot noise. The mean primary electron count received by the APD is assumed large enough to justify the Gaussian approximation. The optical channel is modeled using a log-normal (LN) distribution for low (< 0.75) scintillation indices, whereas for the fully developed speckle case and while implementing single element p-i-n photodiodes, a negative exponential (NE) probability density function (pdf) is used to model the received signal intensity. The array detection assumes an optical channel with correlated random signal intensities across the array. With the aid of simulation, it is shown that the proposed estimators outperform maximum likelihood (ML) estimators for small scintillation indices.; Furthermore, receiver architectures are suggested for optimal combining of an array of avalanche photodiode (APD) detectors in both a binary pulse-position modulated (PPM) scheme and in an on-off-keying (OOK) scheme. Both of these modulation schemes are direct-detection modulation in which the information is carried in the optical signal intensity rather than in the phase or in the frequency of the laser light. Optimum structures motivated by the maximum a posteriori (MAP) rule are suggested. Due to the complexity of the suggested receivers, suboptimum combining schemes are presented. In one scheme, for both binary PPM and OOK signaling, a simple summation of the currents is utilized, while in alternate schemes weighting mechanisms that combine the output of detectors in an optimum or nearly optimum manner are suggested. Simulation results are presented to underscore the superiority of the weighted combining methods as compared to the simple summation algorithms. Analytical performance analyses of the suboptimum detectors are also presented.