Detection, Tracking and Classification of Multiple Targets using Advanced Beamforming and Classification Methods

The problem of detection, classification and localization of multiple ground targets, e.g. trucks and tanks, using arrays of unattended passive acoustic sensors is extremely complicated due to various compounding factors. These include: variability andnon-stationarity of target acoustic signatures, signal attenuation, loss of coherence and fading effects as a function of range and Doppler, effects of wind noise and weather/environmental conditions on the sensory data, competing clutter with similarcharacteristics as those of the targets. In addition, the presence of multiple closely spaced targets can significantly complicate the detection/classification, data association and localization processes, especially when the targets are allowed to remainclose together in tight formations. Consequently, in this project our objectives are to develop, simulate and implement new high-resolution and robust schemes for providing real-time bearing angle estimation and subsequent classification, localization andtracking of different types of narrowband and wideband sources from their acoustic signatures. Furthermore, in developing, simulating and implementing our proposed schemes we will pay particular attention to power consumption and size constraints, which inturn place severe constraints on the complexity of the DSP and communication systems for these arrays. There is a substantial potential for these results to achieve broad-based technology transfer and commercialization capabilities. There are severalgovernmental agencies and companies that are involved in the development of various automatic target detection, classification and tracking systems for different active and passive sensor platforms. The algorithms developed in this project are general andcan be applied to a multitude of similar problems. Therefore, the results of this research could potentially lead to the development of various other acoustic signature analysis systems in a wide market encompassing military and commercial areas. The mainobjective of this Phase II work is to develop, simulate and implement efficient and robust target detection, tracking and classification algorithms for realistic battlefield environments using arbitrary array structures of acoustic sensors. This willprovide not only a solution to the problem of multiple source separation and classification, but also augment the capabilities of the acoustic signature analysis systems presently in use by the Government Labs by providing real-time target/environmentalinformation in realistic operating conditions. The proposed work will also produce a software implementation of all the developed algorithms using MATLAB toolboxes.The potential for this study for a broad-based technology transfer and parternship with other stronger companies is extremely high. There are numerous military, commercial and environmental applications that can benefit from the outcome of the proposedresearch. The development and commercialization of the proposed technologies would lead to the advances in many similar systems covering a wide market potential. Our commercialization strategy would be to identify at least a couple of different potentialapplication areas/markets for the developed technology. The emphasis throughout the proposed research would be on development of algorithms that can easily be implemented on commercially available DSP hardware and COTS components. The implementation ofthese algorithms would also take into account the high throughputs that are required by the targeted real-time applications. Design and implementation of the custom hardware and actual field tests will be addressed in the Phase IIT and Phase II of thisproject,respectively.