J. Cent. South Univ. (2012) 19: 408416 DOI: 10.1007/s1177101210190 Design and compensation of second-order sub-sampling digital frontend WANG Hong-mei(王洪梅) 1 , KIM Jae-hyung 1 , LEE Sang-hyuk 2 , KIM Hyung-jung 3 , KIM Jin-up 3 , KOH Jong-seog 4 1. School of Mechatronics, Changwon National University, Changwon 641-773, Korea; 2. Electrical and Electronic Engineering, Xi’an Jiaotong-Liverpool University, Suzhou 215123, China; 3. Smart Radio Team, Electonics and Telecommunication Research Institute, Daejeon 305-350, Korea; 4. Korea Telecommunication Company, Seongnam 463-771, Korea © Central South University Press and Springer-Verlag Berlin Heidelberg 2012 Abstract: The problem of designing a digital frontend (DFE) was considered which can dynamically access or sense dual bands in any radio frequency (RF) regions without requiring hardware changes. In particular, second-order bandpass sampling (BPS) as a technique that enables to realize the multiband reception function was discussed. In a second-order BPS system, digital reconstruction filters were utilized to eliminate the interferences generated while down converting arbitrarily positioned RF-band signals by using the direct digitization method. However, the inaccuracy in the phase shift or the amplitude mismatch between the two sample streams may cause insufficient rejection of interference. Practical problems were studied, such as performance degradation in signal-to-interference ratio (SIR) and compensation methods to overcome them. In order to demonstrate the second- order BPS as a flexible DFE suitable for software-defined radio (SDR) or cognitive radio (CR), a DFE testbed with a reconfigurable structure was implemented. Moreover, with a view to further demonstrate the proposed compensation algorithms, experimental results show that dual bands are received simultaneously. Key words: bandpass sampling; second order band-pass sampling; reconfiguration; software-defined radio 1 Introduction Wireless terminals of the future are expected to have flexibility, adaptability and cognitivity, which must comply with various coincident wireless standards and should be intelligent to understand the surrounding [1]. Exhibiting these prospects, software-defined radio (SDR), which can realize multiband and multimode operation, has emerged as a next-generation communication technology [2]. Software-defined ratio incorporates the capabilities of wireless terminals in software. By reconfiguring the software, the system can adapt to various wireless standards [3]. In the next-generation SDR receiver, the analog-to-digital converter (ADC) is required to be as close to the antenna as possible, and the frequency conversion and demodulation should be performed using the DSP function [4]. Bandpass sampling (BPS) would be appropriate to realize this. In many recent researches, BPS has been studied for SDR [510]. In this work, in order to design a digital frontend in which ratio frequency (RF) signals in any band can be received and spectrum sensing can be performed without changing the hardware, the second-order BPS signal processing algorithms were studied. In a second-order BPS system, digital reconstruction filters are utilized to eliminate the interferences introduced by down converting the RF bands that are positioned arbitrarily. In the previous studies [1112], a digital reconstruction finite impulse response (FIR) filter for second-order BPS was designed by sampling the available time response obtained from the frequency response, which is the function of the following parameters: sampling rate, time delay between two sampling streams and the frequency position of the input RF signal. However, the inaccuracy in the phase shift or the amplitude mismatch between two sample streams may cause insufficient rejection of interference. Thus, the accuracy requirements for a digital reconstruction FIR need to be analyzed in order to apply it to a practical system. In this work, the degradation of the image rejection performance due to the phase inaccuracy and the amplitude mismatch in the process of reconstruction were formulated. Further, in order to overcome the effect of the errors introduced by BPS hardware, compensation Foundation item: Research financially supported by Changwon National University in 20092010 and the Second Stage of Brain Korea 21 Projects Received date: 20110524; Accepted date: 20111010 Corresponding author: KIM Jae-hyung, Professor, PhD; Tel: +82552133664; E-mail: hyung@changwon.ac.kr