Please use this identifier to cite or link to this item:
|Title:||Signal Processing Techniques for Power Efficiency and Signal Quality Enhancement of SISO and MIMO Radio Systems|
|Keywords:||Engineering--Electronics and Electrical|
|Abstract:||Communications has tremendously evolved during last 30 years period with the goal to realize real-time communications wirelessly. As the communication is targeted to be real-time, there is always a need to achieve the maximum possible data rate. Moreover, link reliability should always be guaranteed to receive a reliable version of the transmitted signal. RF front-ends as the physical layer of the communication systems, suffer from non-ideal behaviour of most of the actual electronic components which causes a nonlinear dynamics. Such systems introduce some amount of distortions to the signal. The other issue which has attracted much attention is the power efficiency which mainly deals with the cost and reliability as well as recently environmental impacts of the communication systems. This dissertation proposes a couple of novel signal processing techniques to overcome the problems associated with the single input single output (SISO) and multiple input multiple output (MIMO) radio systems. The first topic of this research is devoted to efficiently partition the linearization scheme between the base station and mobile terminal. Phase distortions are compensated at the base station transmitter and the compensation of amplitude distortions is devoted to the mobile terminal receiver. This technique improves the power efficiency of communication link and in particular the transmitter. Then, the above technique is extended to base stations to improve the efficiency while meeting the standard spectral requirements. This work employs a soft clipping technique coupled with digital predistortion such that the overall transmitter output spectrum passes the mask. The distortion in the signal amplitude is then compensated at the receiver side.The other major topic of the present research thesis is the dimensionality problem in digital-predistorter design. The information criteria have been employed to consider error along with model complexity to estimate the optimum order. The last research topic carried on in this thesis is related to mitigating numerical instability during system identification of the nonlinear MIMO radio systems suffering from cross-talk. The numerical problem in fixed point processors is resolved using orthogonal memory polynomials. Moreover, a new identification procedure is proposed to reduce computational cost during MIMO digital predistortion identification.|
|Appears in Collections:||Electronic Theses|
Files in This Item:
|ucalgary_2014_vejdaniamiri_mehdi.pdf||3.58 MB||Adobe PDF||View/Open|
Items in The Vault are protected by copyright, with all rights reserved, unless otherwise indicated.