Research On Spectrum Sensing Theories And Techniques For Wireless Communication Systems

As a fastest growing segment of the communications industry, wireless communications have captured the attentions of both academia and industry, and have played an irreplaceable role in people’s daily life. Spectrum is a fundamental support for any wireless communication techniques. However, it suffers from a severe shortage as well as an inadequate utilization, which limits the broad application of wireless communication systems. Cognitive radio, which is seen to be a solution to the contradiction, detects the spectrum holes and provides the opportunities of dynamically access for secondary users. The idea of cognitive radio can also be applied to the femtocell networks to solve the interference coordination issue, thereby increasing the indoor coverage. As an alternative way to improve the spectrum efficiency, OFDM is investigated in this dissertation as well, which is able to support high rate date transmission.On the basis of the existing research works, this thesis conducts theoretical analysis on cognitive radio, spectrum sensing, femtocell, OFDM, and makes some progresses in the fields of wideband spectrum sensing, interference mitigation for cognitive femtocell, parametric channel estimation for OFDM systems.On the topic of spectrum sensing for cognitive radios, a parallel sensing mechanism is put forward. Distinct from the conventional methods in a serial way, the proposal operates over the whole subbands simultaneously and makes the judgments. It firstly estimates the number of occupied channels and then determines the locations. A likelihood function based method is raised, which constructs a series of hypotheses models and turns the issue of estimating the number of subbands into the model selecting problem. The exact locations are then judged by the sample power of each channel. In the next chapter, the smart antenna arrays are adopted in the receiver end. An improved MDL criterion through the multistage Wiener filter structure is put forward, which reduces the computational complexity using the variances of the noise subspace elements rather than the eigenvalues of the covariance matrix. In the meantime, the cross-correlation between the observation data and the reference signals is utilized to suppress the additive noise, which produces cleaner signal and noise subspaces and reduces the probabilities of overestimation as well as underestimation. In another chapter, an algorithm which does not rely on the assumptions of independently and identically distributed Gaussian variables of the primary users’signals is developed. Based on the orthogonality between the signal and noise subspaces, this method separates out the unitary transformed Gerschgorin disks of the signal and noise, and then determines the number of occupied channels using the Gerschgorin disk theorem in a heuristic manner. Therefore, the proposal enjoys a more wide application. Simulations are carried out to demonstrate that these algorithms are robust to noise uncertainty and offer a reliable detection performance.In the aspect of the interference mitigation for cognitive indoor coverage, cognitive femtocell is employed to intelligently change the spectrum usage pattern according to the indoor radio environment. Femtocell is regarded as a secondary system while the macrocell is considered to be a primary system. Via periodic spectrum sensing, femtocell network is coupled with the capability of sensing the surrounding wireless environment, thereby recognizing the interference signals and dynamically access the spectrum holes. In this way, the interference issue is solved efficiently and autonomously. The experiments indicate that the proposed scheme is superior to the existing co-channel and dedicated channel deployment methods.On the topic of parametric channel estimation for OFDM systems, this thesis starts from a novel perspective. The LS criterion is adopted to attain the channel frequency responses on pilot subcarriers. On the basis of the sparse multipath structures, the exponentially embedded family criterion is exploited to estimate the number of channel paths as well as the multipath delays, and then the channel frequency responses on the other subcarriers can be constructed. The numerical results verify that the scheme is able to present an accurate estimation with a low complexity, and thus is suitable for the multipath fading scenario.