In this letter, the cyclostationarity properties of a general class of long-code direct-sequence spread-spectrum (DS-SS) signals are analyzed, and it is shown that these properties can be exploited in long-code DS code-division multiple-access systems. Specifically, it is shown that a continuous-time DS-SS signal exhibits cyclostationarity at cycle frequencies related to both the symbol rate and the chip rate. Moreover, it is pointed out that the discrete-time signal obtained by uniformly sampling a continuous-time long-code DS-SS signal exhibits cyclostationarity, provided that at least two samples per chip are taken. Finally, it is shown how such a cyclostationarity can be suitably exploited for blind signal-parameter estimation

This letter deals with the array detection of a random signal subject to a disturbance modeled as an additive mixture of Gaussian and spherically invariant noise. A detector based on a large sample size approximation of the locally optimum decision statistic is proposed. The implementation of the detection structure requires the knowledge of the correlation matrix of the disturbance at each array element, but is otherwise independent of the noise distribution of the spherically invariant noise component.

The paper deals with two-sensor interception of cyclostationary signals in the presence of additive non-Gaussian noise. The locally optimum approach is considered as a starting point to derive cyclostationarity-exploiting receiver structures for detecting a weak signal and locating its emission source through time-difference-of-arrival (TDOA) measurements. To obtain analytical information about the detection performance of the proposed receivers, the deflection in both constant and variable noise-level environments is evaluated. Monte Carlo simulations, aimed at evaluating the detection performance in terms of detection probability and false-alarm probability, and the TDOA estimation accuracy, have been carried out. The results show that the cyclostationarity-exploiting receivers can significantly outperform the radiometric receivers in both signal detection and source location, when the noise level is unknown or variable and/or strong interfering signals are present.

The letter deals with noncoherent diversity detection of bandpass signals subject to slow and nonselective fading and embedded in correlated non-Gaussian noise modeled as a spherically invariant random process. At first, an asymptotic sufficient statistic for arbitrary fading laws on the diversity branches is derived; then, the asymptotically optimum detector for Rayleigh-distributed amplitude fluctuations is synthesized and its performance is assessed.

Maras and Goutis (IEE Proc. I, Commun., Speech Vision, vol. 136, Aug. 1989, p.233-40) commented on the paper by Izzo and Paura (IEEE Trans. Commun., vol. COM-34, p. 97-103, Feb. 1986). The present paper summarizes the comments of Maras and Goutis and replies to each.

The interception of weak signals in nonGaussian noise is discussed. The spectral correlation property exhibited by all cyclostationary signals is exploited to synthesize multi-cycle and single-cycle detectors which assure a superior tolerance (as compared to radiometric techniques) to one of the most challenging problems in interception, namely accommodating unknown and changing noise level and interference activity. The proposed detectors perform a maximum likelihood estimate of the noise level and use the estimate to form the detection statistic. To obtain some analytical information about the performance of the multi-cycle and single-cycle detectors, the deflection, which is a useful measure of the output (signal-to-noise ratio), especially appropriate for weak-signal SNR interception, has been evaluated. The receiver operating characteristics, carried out via computer simulation, confirm the superiority of the proposed cycle detector with respect to the traditional radiometer.>

The theory of asympoticalllly optimum detection in nonGaussian noise is considered to synthesize a space-diversity structure for the noncoherent detection of a bandpass signal subject to amplitude fluctuations. An asymptotic sufficient statistic is obtained for arbitrary fading laws on the diversity branches and, then, for Rayleigh-distributed amplitude fluctuations, the asymptotically optimum detector is synthesized. The performance of the proposed detection structure is obtained and numerical results and comments are presented. Moreover, a comparison between the asymptotically optimum detector and a simple suboptimum structure is made.

The structure of the locally optimum space-diversity detector for coherent detection of a bandpass signal subject to amplitude fluctuations in the presence of narrow-band non-Gaussian noise is derived and its large-sample performance is evaluated. For easy implementation a suboptimum structure is proposed. It consists of local detectors performing independent local decisions and of a central processor that, working according to simple logical schemes (OR or AND), performs the global decision based on the local ones. Numerical results and a comparison between the performances of both proposed structures are presented and discussed.

The effects of slow and nonselective signal fading on the performance of multilevel noucoherent FSK (NCFSK) systems in an additive mixture of Gaussian and highly impulsive noise are analyzed. For binary systems the bit error rate is derived; forM-ary NCFSK systems upper and lower bounds of the character error probability are obtained. The analysis is performed considering the maximum likelihood receiver for additive white Gaussian noise.