Targeting in Uncertainty (Web Exclusive)

Targeting 1Figure 1: An example Targeting in Uncertainty project output. Tracks labeled 0001SUR14, 0002SUR14, and 0003SUR14 are a simple representation of notional surface trafficking cases. The colored boxes with a red to green color scale give the probability of one or more traffickers at that location in a specified window of time (three hours in this case). Alpha, Beta, and Gamma are surveillance assets (aircraft) and the large blue boxes are recommended search areas for the assets that are optimized based on the amount of contraband likely to be observed.

Dr. Jeffrey G. Morrison

The Joint Inter-Agency Task Force-South (JIATF-S), a subordinate unit within U.S. Southern Command, provides the initial layer of defense against illicit drug traffic from Central and South America. This target-rich setting in the Caribbean and eastern Pacific provides an opportunity to evaluate new technologies against real-world targets in an environment more benign than other naval domains.

 The Targeting in Uncertainty (TiU) project, sponsored by the Office of Naval Research’s (ONR) Warfighter Performance Department, is conducting applied research to address information requirements for JIATF-S in allocating assets for the detection and interdiction of smugglers. The key idea is that by coupling statistical modelling tools with easy to use decision support graphical interfaces, watchstanders can understand what they know and what they don’t know. They then can optimize within this uncertainty and achieve the best possible mission results. The TiU initiative has tested decision-support tools that create a “24/7 virtual targeting board,” which allows watchstanders to evaluate real-time and statistical intelligence, meteorological and oceanographical data, and suspected trafficker behavior.

 The TiU decision-support tools compute the probability of detecting and interdicting potential illicit targets based on current information, allowing watchstanders to assess how best to deploy targeting assets. This project continues to transition these tools to the Fleet Numerical Meteorology and Oceanography Center for ongoing fleet operations, and to integrate them with the Distributed Common Ground System-Navy and Maritime Tactical Command and Control programs for future development and deployment.

The TiU project is a continuation of earlier work building statistically driven decision support. In 2010, ONR and Naval Research Laboratory-Monterey developed the Piracy Attack Risk Surface (PARS) to integrate various types of data to predict which areas would be most susceptible to pirate attack. PARS creates a dynamic decision-support visualization and is updated and published daily in support of Naval Forces Central Command operations. In 2011, ONR, the Space and Naval Warfare Systems Center Pacific, and the Naval Research Laboratory produced a handful of target/case management “widgets” (small web-based applications that can be easily reused in many web-based applications) that will readily transition to the Navy’s Maritime Tactical Command and Control program of record.

 Over the past two years, ONR integrated optimization techniques with existing decision-support tools to produce an asset allocation/decision-support tool for this highly uncertain, asymmetrical targeting environment. JIATF-S asked the ONR team to consolidate related efforts as TiU to directly support the counter-smuggling domain.

 To deliver TiU tools quickly, the project is building on software developed by Sandia National Laboratories that manages position data for current and emerging cases, and allows the TiU tools to be accessible to watchstanders in real time. TiU models now have access to operational data so they can predict the probability of detection and interdiction for specific illicit trafficking targets, as well as for historical traffic flow data. The TiU toolset then uses real-time meteorological/oceanographical-driven models and JIATF-S intelligence data to create probability heat maps (which are graphical representations of data where the individual values contained in a matrix are represented as colors) both for the general flow of illicit traffic conditioned against specific developing cases, selected by operators.  

In November 2013 the TiU team installed an initial asset allocation model incorporating all previous research into a single statistical solution that uses every category of data to characterize and display what is known about developing drug cases, predict activity based on historic data, real-time intelligence, and meteorological factors. The TiU tool will then assist watchstanders in optimizing the allocation of surveillance and interdiction assets to maximize returns for cases across multiple days.

 TiU is assisting both expert and journeyman targeteers at JIATF-S in determining which cases should be targeted for interdiction, given what is known or unknown about all cases in the theater. The tools then provide model-based visualizations that allow watchstanders to optimize the deployment of detection and interdiction assets despite imperfect or incomplete information, and dynamically support current JIATF operational priorities. The tools assist watchstanders by characterizing what is known and unknown about smuggling cases in the theater of operation, and then modeling the effectiveness of detection and interdiction assets against those cases to ensure effective targeting decisions. Initial asset allocation recommendations can now be generated on demand, in five to 10 minutes, compared to the hours required previously.

 An example guidance product is shown in Figure 1. Three notional cases are presented along with three notional surveillance assets (Alpha, Beta, and Gamma). JIATF-S intelligence analysts provide available information about the distribution of departure locations, departure times, waypoint locations, and arrival locations, along with information about suspected speeds, conveyances, loads, and behaviors. That information is combined with environmental factors such as wind, waves, currents, and time of day that are computed in a suite of statistical models and result in gridded probability surfaces.

Targeting 2Figure 2: JIATF-South’s concept for a “Virtual 24/7 Targeting Board”

 These statistical surfaces are the basis for the heat maps that communicate the relative probability of one or more illicit traffickers being at a particular location during a particular time period (shown as the colored squares). JIATF-S operators have complete freedom in the specification of the factors that drive the probability model, enabling them to effectively perform sensitivity experiments to better understand what information will have the greatest impact on the predicted results. This allows operators to produce solutions that reflect their assessment of specific cases as well as adapt to current JIATF-S operational objectives.

 The TiU system also provides asset allocation guidance. Friendly force information is pulled from JIATF-S that includes aircraft type, location, speeds, endurance, crew-rest cycle, scheduled departure times or windows, sensor assets, and the performance of those assets as a function of target vessel and environment. The operator can select the time period over which to optimize (e.g., 24 or 48 hours) and the objective function (such as total weight of drugs).

 An example of a targeting plan is shown in Figure 2 for a 48-hour optimization that aims to maximize contraband weight. Search boxes are provided along with the time at which the aircraft should depart, the time at which the box should be searched, and the duration of the search. Note that because of the 48-hour window the Alpha asset is able to make two flights. Constraints such as crew rest cycles prohibit Beta and Gamma from making a second flight in this instance.

 Efforts continue to refine these capabilities at JIATF-S and to migrate this approach to managing uncertainty to other warfare domains. The tools developed also are ready to deliver to Navy command and control and intelligence programs of record.

About the author:

Dr. Morrison is a program officer with the Office of Naval Research.