Information-theoretic signal design, coding techniques and processing algorithms for high resolution delay-Doppler radar
Abstract
In this thesis, we investigate coding techniques and processing algorithms for the radar resolution problem. In the first part of the thesis, we design signals which maximize the mutual information (MI) between the received signal and target impulse response in a Bayes decision-theoretic approach. Because of the problem's analytical intractability, we set up and solve an approximate problem. In order to improve the derived solution, a numerical algorithm is proposed to design MI-maximizing signals which are shown to provide a water-filling solution. It is shown that when compared with SNR-maximizing signals, MI-maximizing signals provide better detection performance for less probable targets because of their water-filling type energy scheduling. In the second part, we design phase modulated space-time codes for MIMO radar for high delay-Doppler resolution. A systematic approach is adopted where necessary and sufficient conditions are derived for perfect delay-Doppler sidelobe cancellation and a design approach is proposed to construct codes satisfying these conditions. Such codes therefore produce a composite ambiguity functionwhich is very close to ideal thumbtack. We also propose a processing algorithm for our model and provide simulation results in the presence of multiple scatterers to show that the designed codes along with the proposed processing algorithm produce a high resolution delay-Doppler image and therefore make unambiguous delay-Doppler estimation possible. Finally, we consider a multi-carrier MIMO radar model and propose non-linear processing schemes which provide far better resolution performance than standard matched filtering. Since our processing schemes are not optimal from detection point of view, we investigate the degradation in detection performance when compared to the matched filter and show that our proposed schemes offer significant improvement in resolution at the expense of minimal SNR loss.
Degree
Ph.D.
Advisors
Bell, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.