Waveform-agile sensing for target tracking and detection in clutter

Recent advances in electromagnetics, processing, and radio frequency analog devices afford a new flexibility in waveform design and application. This technology is now being used in the research and development of sensors that can dynamically change the transmit or receive waveform in order to extract information that can contribute optimally to the overall sensing objective. Thus, a new paradigm of waveform-agile sensor processing has been exposed, which promises significant improvements in active sensor system performance by dynamic adaptation to the environment, target, or information to be extracted. This dissertation considers two such problems, for which it develops waveform-agile sensing algorithms for tracking and detection of targets in clutter.; The first problem involves the dynamic selection and configuration of waveforms for two active sensors that track one or more targets moving in two dimensions. Within the context of a realistic, nonlinear observations model, the wave form-scheduling algorithm chooses the phase function, duration, and frequency modulation rate of the next transmitted waveform so as to minimize the mean square tracking error. The waveforms belong to a class of generalized frequency-modulated chirps with linear or nonlinear time-frequency signatures. Two different approaches are explored for the prediction of the mean square tracking error, based on the Cramer-Rao lower bound on the measurement error variance.; The second problem develops a waveform-agile detection algorithm that aims to dynamically design the next transmitted waveform to improve the detection of small targets on the ocean surface in the presence of heavy sea clutter. This is a challenging problem in radar applications. Using the well-established compound-Gaussian model for sea clutter, a subspace-based method to mitigate the clutter is developed, which exploits the statistical differences between clutter and target returns. Maximum-likelihood estimates of the clutter statistics are then employed in a mean square optimization approach to dynamically design a phase-modulated waveform for the next transmission. A simulation study based on parameters extracted from real sea clutter data is presented to demonstrate the benefits of this approach.