Reduced complexity variable precision signal processing for digital communications

As digital cellular communications transforms personal communication and its interaction within society increases, the complexity of the communications system increases to counter variable SNR environments. One method to reduce system complexity is to replace computationally expensive components with low-complexity ones. By replacing the conventional signal autocorrelation, with combined reduced complexity correlation estimator and variable bit quantization, considerable computational savings are achieved. The accuracy of these algorithms is of concern especially in applications, which are resolution sensitive such as in the direction of arrival estimate produced by beamforming. Simulations show that the reduced complexity quantized correlation estimators can achieve comparable results to the convention method. For the beampattern the quadruplex using 16 and 8 bit wordlengths showed similar beamwidths and were within 3% for 16 bits and 10% for 8 bit wordlength DOA compared to the conventional method full floating point.