Wide Area Persistent Surveillance (WAPS)

Wide Area Persistent Surveillance is defined as the ability to provide surveillance over as much of the region known to be associated with a specific activity in order to increase the chance of detecting and observing the activity, identify the entity, track the entity forward in real time or backwards forensically. In most cases, the activity of interest can be identified as a transaction between locations.

Automatic tracking algorithms perform poorly on real data. Performance of pattern analysis is untested for asymmetric warfare. PowerPoint products are time consuming to produce and not easy to understand. There is no formal training and only a limited numbers of WAPS analysts. There is inadequate exploitation tradecraft, tools, and automation to leverage the inherent value of collected WAPS datasets in a timely fashion.

CTTSO is interested in proposals that address WAPS exploitation challenges. The intent of the advertised requirements is to identify technologies and approaches that provide near-, mid-, and long-term solutions that enhance the capabilities of the U.S. Government to combat or mitigate terrorism. The main objective is to provide innovative approaches, tools and techniques for improving WAPS exploitation and analysis capabilities that can be rapidly transitioned into operational use.

The WAPS enterprise is expanding to include data from many sensors and sources. Electro-Optical (EO) motion imagery from Constant Hawk and Light Hawk, and Ground Moving Target Indicator (GMTI) data from Joint Surveillance Target Acquisition Radar System (JSTARS) and Littoral Surveillance Radar System (LSRS) are the primary data sources of interest. The volume and complexity of WAPS data represents a challenge for human analysts to manage large datasets, integrate data from multiple sources, identify activity of interest, isolate it from background activity, and derive useful intelligence in a timely manner.

The Government previously conducted the BlueGrass experiment to build data sets representative of real world WAPS challenges. The BlueGrass dataset contains coincident coverage from two GMTI sensors (JSTARS and LSRS), and two wide area EO sensors (Constant Hawk and Light Hawk, the later of which remains classified). A variety of scripted, operationally relevant, activities and behaviors were executed in support of this effort. The data contains a wide range of both naturally occurring (illumination, cloud cover, cloud shadow, haze) and man-made (traffic density, occlusion) environmental conditions and backgrounds in urban, suburban and rural settings. BlueGrass data sets will be provided to submitters selected to submit Full Proposals as the basis for developing solutions to the requirements described below. Prior to receiving data sets, proposers are required to sign privacy and nondisclosure statements. EO sensor data is FOUO; however GMTI data is classified SECRET and will be restricted to those proposers with active personnel and facility clearances.

Example cases for activity detection and challenge problems derived from the BlueGrass data will be provided including varying degrees of detection difficulty due to target type/size, background contrast, environmental and background conditions, and image quality. A diversity of sun angle, occlusion, shadowing, traffic density and landscape (urban, suburban, rural) conditions will be included to the greatest extent possible. Data sets will be customized to include a subset of analyst vetted human activities representing the challenge and include:

• a specified geographical area (bounding box);
• EO and GMTI data (formats including natoex, stanag4670, JPEG 2000 (JP2), and shape files) for the time range of the example; and
• limited GPS "ground truth" data of vehicle and personnel movements within the specified geographical areas during the time of interest.

Additional BlueGrass data sets will be withheld and used by the Government for independent evaluation and verification of developed capabilities and approaches, which shall be delivered in a form that allows the Government to replicate contractor test results.

Fundamentally new strategies and methodologies to leverage wide area EO and GMTI data to detect, characterize, track and analyze moving entities and patterns of behavior are sought. The key interest and focus is on improving performance of WAPS analysts. In particular, efforts shall focus on:

• software automation tools to improve user experience and productivity;
• advanced hardware to accelerate data playback and visualization;
• algorithms for automated detection, tracking, characterization and alerting to activity of interest in WAPS datasets;
• enhanced methods for visualization of WAPS datasets;
• product production and storyboarding methods; and
• tasking workflow and analysis task management.

Hardware, software, algorithmic, preprocessing, enhanced visualization, and other novel solutions may be applied individually or in conjunction with new or existing WAPS applications and infrastructure. Component capabilities that characterize discrete activity of individual entities should allow for aggregation to represent complex interactions, behaviors, associations and hierarchies of groups of entities within or across data types. Algorithms shall discriminate between human generated effects and environmental or processing artifacts (such as wind-blown objects, solar reflections, processing effects, storms and terrain effects) with the goal of developing the ability to characterize regular movements, patterns and schedules.

Proposals shall address the ability to perform functions automatically, semi-automatically or provide tools for humans to detect, identify, track and analyze activity in persistent surveillance data sets. Successful capabilities will increase the overall efficiency, accuracy and yield of intelligence information derived from WAPS datasets. Current human analyst exploitation efficiency and accuracy are the baseline to which improvements will be compared. Proposals are sought to address four general research development, test, and evaluation (RDT&E) thrusts. Proposals need not address every requirement in the four general RDT&E requirements, but rather may include innovative proposals for one or more individual elements in the requirement.