Equalization in a Direct Sequence Spread Spectrum System using a Least Squares Lattice Filter

A direct sequence spread-spectrum system, employing a least-squares equalizer is shown to perform nearly as well as the optimal receiver under certain conditions. The system presented here requires neither the knowledge of the noise statistics and of the impulse response of the channel, nor a training sequence of known data bits. The reference for the equalizer is provided by the decoding circuit with negligible delay. Simulation results for the probability of error are close to the theoretical results of an optimal receiver. To adapt to changes in noise statistics and channel, a restriction on the memory of the equalizer is required, which causes some degradation in performance.