Adaptive Sensing and Target Tracking for MIMO Radar Systems
The objective of this research project is to identify algorithms that leverage recent advances in adaptive sensing for application to Multiple-Input Multiple-Output (MIMO) radar systems. The term adaptive sensing refers to sensing modalities in active Intelligence, Surveillance, and Reconnaissance (ISR) systems that employ feedback to select transmitted waveforms in response to previously processed data, to adaptively optimize overall system performance. MIMO radar systems are those systems with multiple transmit and receive apertures, and the capability for independent signal selection at each transmit aperture. We use adaptive sensing for the high-dimensional problem of multiple signal selection in MIMO radar systems. A complementary problem is that of target tracking based on the selected transmit signals and the optimal statistical processing of the data collected at the receive apertures. Significant improvements in tracking performance using the proposed approach are expected and will be quantified.
The specific objectives of this research are as follows. We first identify algorithms for adaptive sensing and target tracking under a simple point target model. This includes the generalization of existing adaptive sensing results to the adaptive sensing of state vectors with random amplitude, then target tracking using a known propagation channel response, and finally target tracking for an unknown propagation channel response. Second, we are developing models for what we have termed complex point targets,and based on these models we will develop methods for statistical inference on these targets, including detection, parameter estimation, and tracking. Third, we will investigate the effects of uncertain local oscillator (LO) synchronization across all the transmit and receive apertures in a MIMO radar system, which is inevitable given the spatial distribution of the apertures. Methods for mitigating these effects, that can be included in the adaptive sensing and statistical inference algorithms, are being derived. Finally, we will integrate the results of these three efforts just described, leading to algorithms for tracking complex point targets using a MIMO radar system and adaptive sensing methodology, taking into account system synchronization issues.
Compression and Cooperation for Wideband Spectrum Sensing and Cognition
The objective of this research is to increase the detection resolution and reduce the implementation cost of wideband spectrum sensing and cognition, which are challenging core issues in spectrum-sharing cognitive networks. The approach seeks to leverage the benefits of compressive sampling and user cooperation to develop a collaborative compressed sensing framework for wideband networks. Research thrusts include compressed wideband sensing at affordable signal acquisition costs and reliable collaborative sensing with low network overhead.
This research combines compression and collaboration techniques with technical innovations in wideband spectrum sensing. The project seeks to develop a suite of compressed sensing algorithms that exploit spectral sparsity to reduce the need for high-speed sampling in ultra-wideband and wideband radios. A decentralized approach for joint cooperation and compression, which exploits spatial diversity to alleviate the hidden terminal problem caused by wireless channel fading, is investigated for ad hoc cognitive networks. Further, the research examines fundamental tradeoffs in a cooperative sensing system including diversity gain, compression, sensing time and complexity.
This project will lead to new techniques that can improve the spectrum utilization efficiency and user capacity of wireless networks, thus allowing a multitude of new cognitive radio transceivers. Such devices include wireless sensors and radio frequency identification (RFID) chipsets for monitoring and tracking applications. The compressed sensing and decentralized collaboration techniques being investigated also have the potential to contribute to other sensing-related applications, such as wireless sensor networks. Research experiences gained through this project help to prepare students for engineering in the 21st Century.