Multipleinput multipleoutput (MIMO) radar employs orthogonal or partially correlated transmit signals to achieve performance benefits over its phasedarray counterpart. It has been shown that MIMO radar can achieve greater spatial resolution, improved signaltonoise ratio (SNR) and target localization, and greater clutter resolution using spacetime adaptive processing (STAP). This thesis explores various methods to improve the signaltointerferenceplusnoise ratio (SINR) via transmit and receive beamforming.
In MIMO radar settings, it is often desirable to transmit power only to a given location or set of locations defined by a beampattern. Current methods involve a two step process of designing the transmit covariance matrix R via iterative solutions and then using R to generate waveforms that fulfill practical constraints such as having a constantenvelope or drawing from a finite alphabet. In this document, a closed form method to design R is proposed that utilizes the discrete Fourier transform (DFT) coefficients and Toeplitz matrices. The resulting covariance matrix fulfills the practical constraints such as positive semidefiniteness and the uniform elemental power constraint and provides performance similar to that of iterative methods, which require a much greater computation time. Next, a transmit architecture is presented
â€¨that exploits the orthogonality of frequencies at discrete DFT values to transmit a
sum of orthogonal signals from each antenna. The resulting waveforms provide a lower meansquare error than current methods at a much lower computational cost, and a simulated detection scenario demonstrates the performance advantages achieved.
It is also desirable to receive signal power only from a given set of directions defined by a beampattern. In a later chapter of this document, the problem of receive beampattern matching is formulated and three solutions to this problem are demonstrated. We show that partitioning the received data vector into subvectors and then multiplying each subvector with its corresponding weight vector can improve performance and reduce the length of the data vector. Simulation results show that all methods are capable of matching a desired beampattern. Signaltointerference plusnoise ratio (SINR) calculations demonstrate a significant improvement over the unaltered MIMO case.
Date of Award  May 12 2013 

Original language  English (US) 

Awarding Institution   Computer, Electrical and Mathematical Science and Engineering


Supervisor  MohamedSlim Alouini (Supervisor) 

 MIMO Radar
 Waveform Design
 Beamforming
 DFT