Dynamic precision scaling for low power WCDMA receiver

One of the most important applications of digital signal processing (DSP) is wireless communication. This kind of application requires low power implementation of DSP, which generally uses fixed-point arithmetic. The fixed-point architectures should be developed to maintain the energy consumption power at a reasonable level. In this paper, an approach which adapts the fixed-point specification according to the input receiver signal-to-noise ratio (SNR) is proposed. To underline our approach interest, the rake receiver of a WCDMA receiver is examined. Results show about 25% - 40% energy savings with our dynamic precision approach.