Performance of direct sequence spread spectrum receivers in pulsed noise jamming and in narrowband interference

We consider the detection of a direct sequence binary phase-shift keyed (DS/BPSK) signal in two types of jamming environments: pulsed broadband noise jamming and narrowband interference.; Optimal receiver structure for direct sequence signal received in pulsed broadband Gaussian noise jamming requires the side information regarding whether the jammer is on or off. The Expectation-Maximization (EM) Algorithm is used to estimate the jammer state and the jammer parameters and obtain a decision on the binary signal based on the estimated likelihood functions. The performance of this receiver is compared with those of a maximum-likelihood receiver, an optimal (unrealizable) receiver, linear receiver, hard-limiter and a soft-limiter.; Maximum likelihood receivers for direct sequence signal in Gaussian narrowband interference are derived. The maximum likelihood receiver that uses only the observations from the interval of the bit that it is detecting is linear. The maximum likelihood receiver that uses observations from the previous bit interval and its bit error rate are derived. This receiver is nonlinear in observations. The performances of the two maximum likelihood receivers are compared to those of transversal filters. The nonlinear maximum likelihood receiver outperforms the linear maximum likelihood receiver and the one-sided and two-sided transversal filters. For transversal filters, new error rate expressions that account for intersymbol interference effects are derived.; The suitability of SNR as a measure of bit error rate (BER) performances of direct sequence receivers employing interference rejection filters is studied. Kalman filter, a linear modification and a nonlinear modification are studied by simulation. In the case of small processing gain, short PN sequences, low thermal noise and strong interference, the output SNR of the nonlinear filter is higher than those of the linear filters, but the nonlinear filter also exhibits a higher error rate. For moderate processing gain, SNR provides reasonably accurate error estimates for all the filters.