Spectrum sensing algorithms and transceiver systems design for cognitive radio network testbed

Cognitive radio has become the breakthrough wireless technology in the past decade. It is a network that can use licensed bands for communications, whenever it would not cause any interference (by avoiding them whenever legitimate user presence is sensed). Users in the network are fully programmable wireless devices. They should be able to sense their environment and dynamically adapt their transmission waveforms, channel access methods, and networking protocols. The research of cognitive radio is multidisciplinary. In this dissertation, we study on two key enabling technologies: Spectrum sensing algorithms and transceiver systems design.;Spectrum sensing detects the unused frequency bands. Two paradigms are considered: Single band spectrum sensing and wideband spectrum sensing. In single band spectrum sensing, the unlicensed user detects the availability of one frequency band at each sensing operation. We propose to model this problem as a low rank signal detection problem. Discrete Karhunen--Lo{`e}ve transform and general likelihood ratio tests will be used to derive new algorithms under the low rank signal detection model. Simulation results, hardware implementations and real-world experimental results will illustrate the performance improvement of the new algorithms. In wideband spectrum sensing, the unlicensed user is able to detect the availability of multiple frequency bands at each sensing operation. Sensing with feasible sampling rate and acceptable computation complexity are very challenging. We will analyze the practical properties of wideband spectrum sensing and summarize the features that can assist developing efficient wideband spectrum sensing algorithms. Compressed sensing theory will be used extensively and modified compressed sensing algorithms will be proposed.;Transceiver systems design studies the programmable high performance platform for cognitive radio devices. General architecture for cognitive radio will be introduced. Two types of cognitive radio systems are considered: Overlay systems and underlay systems. We will introduce several state-of-the-art platforms that can be used for the overlay systems. As for the underlay systems, we will use convex optimization theory to assist the systems design, featuring low complexity with optimum programmable waveforms. Time reversal waveforms used at the transmitter is a special case of cognitive radio that consists of intelligent programmable waveforms. Two generations of time reversal systems implementations will be introduced. Compressed sensing theory will be introduced to design the underlay systems with sampling rate that is far less than Nyquist sampling rate. Relative ultra-wideband channel estimation sub-systems with sub-Nyquist sampling rate will be introduced.