Software Radio, Digital Signal Processing And Parameter Estimation In Mind

With the very rapid advancement of wireless communications, a new challenge to this industry is the integration of multiple systems and applications on a single device. Software Radio, firstly introduced by Jeo Mitola in 1992, addresses the goal of developing transceivers that will operate with several standards and in several frequency bands on a common hardware platform. Such a platform would allow flexible and programmable transceiver operations. This type of software radio is expected to be a key technology in several application scenarios of wireless communications, including cellular telephony, military communications, aeronautic data link and etc.The practical structure of the hardware platform of software radio is designed to cope with the linear modulation schemes, such as MPSK, OQPSK, QAM and MSK. In this thesis, we introduce a linear software radio architecture that can be extended to handle GMSK, which is a nonllnear modulation scheme. The synchronization algorithm and the design of the matched filter of this architecture suitable for GMSK signals are also studied.Synchronization is very important in digital receivers. A new non data-aided (NDA) and all-digital algorithm is proposed for join estimation of timing and carrier phase in GMSK modulation. It is based on maximum-likelihood arguments and has a feed forward structure. The proposed algorithm is also suitable for linear software radio. Performance in AWGN channel is assessed by computer simulation. and is compared with the modified Cramer-Rao bound and other existing algorithms. Simulation results have shown that the performance of the proposed algorithm is close to MCRB and other algorithms. Moreover, the proposed algorithm can work at an over sampling rate as low as 2, implying wider bandwidth and higher transmission rate, or lower hardware resource and power consumption.Cyclostationarity (CS) method is very useful in signal processing for communications. Frequency offset and symbol timing recovery in flat-fading channels based on cyclostationarity was introduced by Gini and Giannakis, and this method has many advantages. In this thesis, we will show that under some additional assumptions, the CS method can be extended to frequency-selective fading channels. Simulations have proved the feasibility of this extension, with only a small degraded performance.Finally, the implementation of the hardware platform of software radio for aeronautic data link is introduced.