| Locally optimum (LO) detection is an appropriate technique for removing strong interference signals because its performance asymptotically approaches the global optimum as the signal amplitude goes to zero. However, deriving a LO detector with memory for correlated interference is often practically impossible due to the need to deal with multi-variate probability density functions (pdf). Assuming Markovian structure for the interference and subsequently applying an autoregressive model leads to a straightforward implementation of the LO detector.
Intentional jamming and other forms of correlated interference are common problems plaguing spread spectrum systems. This paper presents a LO detector based technique to reliably detect spread spectrum signal under strong interference conditions. The LO detector involves computing a nonlinearity based on the pdf of the interference. The nonlinearity is computed for the received signal values and a large number of these values, equal to the spreading gain of the system, are combined together to produce the decision variable. This is then compared against a fixed threshold depending on the type of hypothesis test used.
Performance of the LO detector is studied by performing Monte-Carlo simulations. Spread spectrum signal corrupted by strong correlated interference is simulated and this is then passed to the LO detector where an estimate of the transmitted information is formed. The overall bit error rate (BER) of this detector is studied for different kinds of interferences, different interference levels and other system parameters. Finally, the performance of the LO detector is compared to that of some other commonly used interference suppression techniques and the conditions under which the LO detector works better than traditional methods are highlighted. |