Oversampled Hybrid Filter Banks for High-Dynamic-Range Cognitive-Radio/Software-Defined-Radio Receivers

An oversampled Hybrid-Filter-Bank (HFB) receiver architecture for Cognitive-Radio (CR) and Software-Defined-Radio (SDR) applications is presented in this paper. With this novel receiver configuration, improved dynamic range in relation to that of a conventional single-band receiver is attainable even when using lower-resolution analog-to-digital converters (ADCs). To verify the proposed approach, a proof-of-concept intermediate frequency (IF) receiver prototype consisting of a five-channel HFB is developed and tested. Specifically, signal-reconstruction experiments for QPSK-modulated signals with and without the presence of an interfering signal are conducted to show that this HFB receiver outperforms its equivalent single-ADC counterpart.

1. Introduction

Cognitive Radio (CR) and Software-Defined Radio (SDR) are emerging concepts in the future wireless communications scenario. Advanced features such as dynamic spectrum access, spectrum sensing, system reconfiguration and interoperability will offer modern radio systems more efficient ways to exploit the electromagnetic spectrum [1]. However, to ensure these figures of merit, their RF front-end must be capable of sensing ultra-broad-band spectral ranges while dealing with high and low-power signals at the same time. This leads to very demanding ADC specifications, where elevated sampling rates and dynamic ranges are essential. Besides, the ADCs must be reasonably priced and have low DC-power consumption. These requirements can not be met with current technology.