MATCHED WAVEFORM DESIGN AND ADAPTIVE BEAMSTEERING IN COGNITIVE RADAR APPLICATIONS

Abstract

Cognitive Radar (CR) is a paradigm shift from a traditional radar system in that previous knowledge and current measurements obtained from the radar channel are used to form a probabilistic understanding of its environment. Moreover, CR incorporates this probabilistic knowledge into its task priorities to form illumination and probing strategies thereby rendering it a closed-loop system. Depending on the hardware's capabilities and limitations, there are various degrees of freedom that a CR may utilize. Here we will concentrate on two: temporal, where it is manifested in adaptive waveform design; and spatial, where adaptive beamsteering is used for search-and-track functions. This work is divided into three parts. First, comprehensive theory of SNR and mutual information (MI) matched waveform design in signal-dependent interference is presented. Second, these waveforms are used in a closed-loop radar platform performing target discrimination and target class identification, where the extended targets are either deterministic or stochastic. The CR's probabilistic understanding is updated via the Bayesian framework. Lastly, we propose a multiplatform CR network for integrated search-and-track application. The two radar platforms cooperate in developing a four-dimensional probabilistic understanding of the channel. The two radars also cooperate in forming dynamic spatial illumination strategy, where beamsteering is matched to the channel uncertainty to perform the search function. Once a target is detected and a track is initiated, track information is integrated into the beamsteering strategy as part of CR's task prioritization.