Home Defence & Homeland Security JASON reports on ”compressive sensing” for DoD sensors

JASON reports on ”compressive sensing” for DoD sensors

US: The elite JASON scientific advisory panel released its latest report, ”Compressive Sensing for DoD Sensor Systems”. During its 2012 Summer Study, JASON was asked by ASDR&E (Assistant Secretary of Defense for Research and Engineering) to consider how compressed sensing may be applied to Department of Defense (DoD) systems, emphasising radar because installations on small platforms can have duty cycles limited by average transmit power.

JASON is an independent scientific advisory group that provides consulting services to the US government on matters of defence science and technology. JASON typically performs most of its work during an annual summer study, and has conducted studies under contract to the Department of Defense (frequently DARPA and the US Navy), the Department of Energy, the US Intelligence Community and the FBI. Approximately half of the resulting JASON reports are unclassified.

According to the report, the ”Compressive or sparse sensing” represents a conceptual approach for enhancing the capabilities of DoD sensor systems used for image generation. Many DoD sensor systems support multiple functions (e.g., multi-mode radar performing both surveillance and SAR) which often compete for the sensors resources (e.g., dwell times, beam positions). Other sensors generate huge volumes of data (e.g., airborne/overhead EO/IR) which can overwhelm communications links utilised to send this information to users at other facilities.

In other instances, operators may want sensors with large physical apertures to achieve good angular resolution but cannot afford to fully populate the entire array with sensor elements due to cost, power and/or weight considerations. (In other words, compressive sensing can lead to improved array angular resolution performance whereby a larger array with the same number of elements as the original smaller physical aperture array are arranged in a pseudo-random pattern resulting in improved angular resolution with the application of compressed sensing.) All of these situations represent potential candidates for a relatively new technology approach known as compressive sensing.

Some of the recommendations of the report include:
1) DoD can and should play a major role in exploring where and how compressed sensing can be applied, particularly to radar and optical systems. These efforts should include applying new sparse reconstruction algorithms to old deconvolution problems as well exploring new systems.

2) To find where and how ”Compressive Sensing’ can benefit DoD radar applications, DoD should develop a strongly guided program of 6.1/6.2 research to (a) develop a sparsity library for important types of targets, (b) quantify how CS degrades target identification through Receiver Operating Characteristic (ROC) curves, (c) create performance metrics for evaluating reconstructed signals, (d) develop operational experience with CS-radar with test beds on different types of radars, and (e) perform regular reviews and provide guidance from people experienced in military radars.

3) If attractive ”Compressive Sensing’ radar applications are found, they should be developed in conjunction with software-defined, cognitive radars to provide the needed flexibility in choosing when and how sparse illumination is used.

4) Although this is not necessarily an example of compressed sensing, DoD should consider consolidating GMTI (Ground moving target indicator) and SAR (Synthetic aperture radar) functions in a ‘Foveal Radar’ that subdivides the coherent processing interval to obtain coarse identification of movers and then switches to full SAR for high-resolution images. Pulses are neither skipped in this mode, nor is resolution compromised in the final images.

Source: FAS