Advancements in Active Shooter Simulations
Improved modeling and simulation methods are now allowing for detailed virtual investigation of active shooter scenarios ahead of an attack.
By Larry M. Bryant, Ph.D., M.SAME, and Kenneth W. Herrle, P.E., CPP, M.SAME
Recent large-scale massacres in Colorado (2012), Mumbai (2008) and at Virginia Tech (2007) underscore the need to better understand, plan and train for previously unconsidered active shooter attacks on an unsuspecting and vulnerable population. Yet, effective planning for such events historically has been constrained by the limitations of available methods and tools. That is changing.
Within the past several years, even guarded facilities such as stateside military installations have become a target for these horrors, evidenced by the 2009 Fort Hood attack, and the foiled plots at Fort Dix (2007) and again at Fort Hood (2011). Terrorist organizations and lone individuals possess the capability to attack these guarded installations, in addition to far more vulnerable establishments such as schools, stadiums, transit centers and other public gathering places. Until recently, it just was not possible to obtain detailed post-mortem style information for these scenarios prior to an actual attack. Now, however, new advancements in modeling and simulation methods are allowing for detailed virtual investigation of such scenarios beforehand. This can mean highly effective mitigation, response planning and training.
AGENT-BASED MODELING AND SIMULATION
In an instant, terror events spread an environment of chaos. There is urgent need to effectively and efficiently remove people from harm’s way, particularly during prolonged active shooter engagements—like the 2009 Fort Hood attack. To truly prepare for such events, advanced modeling and simulation becomes essential in planning effective response and recovery operations due to the inherent difficulty of obtaining realistic and useful data by other means.
When responding to these scenarios, there is much more to consider than simply deploying security forces, evacuating individuals through the nearest exits, or blindly sheltering in-place. It also is critical to know what the next probable step of the aggressor(s) is and what is the most effective method of neutralizing the situation at hand. Furthermore, planning should consider the number of potential casualties (injuries) and fatalities that may occur. It should ascertain the number of security forces, first responders and ambulances required for effective response; the estimated quantity of hospital beds and triage medical staff required for treating casualties; and the realistic availability of resources for meeting each of these needs in “real time.”
Detailed data for these complex and random events is virtually impossible to obtain solely through routine drills, traditional tabletop exercises, typical flow-based pedestrian modeling, or historic precedence.
Until recently, analytic methods for modeling and simulating mass-casualty events (such as active shooter) have lacked sufficient fidelity and realism to deliver successful evaluations. A popular model typically used for mass movement of individuals during an evacuation is the hydraulic model, which essentially treats the mass movement of people as a flowing fluid. However, these fluid-based models reveal flaws when it comes to realistically replicating pedestrian evacuations during an event—including the assumption that all people will start moving at the same time, and that everyone will move uniformly in both speed and direction. Such models also do not account for the differences in human behavior, emotional states, or cognitive decision-making processes. They do not presuppose personal knowledge of the facility, availability of response resources, or any high-profile status of select individuals.
Advances in state-of-the-art modeling techniques and cognitive task analysis science are facilitating the development of advanced agent-based modeling and simulation applications. These agent-based applications represent the next-generation of mass-casualty modeling and simulation analysis. Each analyzed individual is recognized as its own unique autonomous agent with its own distinctive set of characteristics and knowledge base.
By using agent-based modeling, planners can quickly assess a range of attack scenarios with unique attributes for the aggressors, bystanders, responders, facility and/or installation, and specialized targets within the facility/installation. Another unique aspect to this type of modeling is that individual agents practice situational awareness. They are able to share with other agents they encounter more specific knowledge about the facility/installation or event in question.
Development of these applications requires extensive involvement of software developers, mathematicians and various other subject matter experts. It also demands cognitive psychologists who can address idiosyncrasies in human behavior and cognitive decision-making ability. In addition to decision-making based on deductive reasoning, past experience and specialized knowledge, such models also must consider a range of psychological parameters. These can include mental states when people are suddenly awakened, disoriented, frightened, or panicked. Or they could be mental distractions such as parents evacuating with small children. Routine habits also can be captured in agent-based models. For example, people typically exit through the door they came in, even if there is a closer or more convenient exit nearby.
New agent-based modeling and simulation tools—such as the event simulation tool E-Sim—provide the ability to maintain the fidelity of the specific facility that’s being modeled. These models are realistic, dynamic and replete with the capabilities to analyze a full-suite of adverse events, both natural and manmade. [Fig. 1 shows an E-Sim agent-based simulation of a large-scale event in a high-rise hotel, illustrating the level of fidelity that can be captured using such models.]
By using advanced agent-based modeling with cognitive task analysis, it is possible to analyze the progression of assailants through a facility or installation in any number of hypothetical adverse situations. The objective is to glean critical information from the event and assess multiple possible outcomes based on the effectiveness of response to the event. Such information may include:
- Time/rate of progression of the assailants through the facility or installation-based on aggressor experience/training, capabilities, knowledge of the facilities and procedures, and planning of the attack.
- Location and vulnerability of potential targets within the facility or installation.
- Optimal avenues of approach for security response forces based on facility/installation layout, situational details, experience/training and capabilities.
- Anticipated interception/neutralization time required by response forces.
Fig. 1: E-Sim advanced simulation of large-scale event in a high-rise hotel
Fig. 2: E-Sim advanced simulation of an active-shooter attack.
- Potential “friendly-fire” incidents based on response force positions and avenues of approach.
- Optimal evacuation avenues (or shelter in-place positioning) for facility occupants.
- Optimal communication and annunciation based on situation and areas affected.
- Obstacles presented by the facility/installation or event—such as unfamiliarity with the building layout; smoke or darkness; locked or hardened doors and windows; hostile fire areas; and damaged or impassable vehicle or pedestrian corridors.
- Delays in facility occupant response to the event such as injury, uncertainty, fear, denial, disability or separation from other occupants.
- Anticipated number and location of casualties/fatalities.
- Number of first responders, medical staff, ambulances and hospital beds required for response and treatment of casualties.
- Added benefits provided by directed evacuation or shelter-in-place of bystanders and shared knowledge throughout the event.
Although this is not an all-inclusive list, these provisions address critical topics that will aid in ensuring effective preparation for response and mitigation of potentially catastrophic attacks.
Initial development of advanced agent-based simulation for active shooter events has resulted in a specialized analysis tool that accounts for much of the information described. [An example of select analysis output is depicted in Fig. 2, which shows, within a relatively simple building layout, a specific point in time of the attack and the response of three aggressors and three defenders.] The time-based actions and reactions address facility layout, defensive positions, protected assets, breaching of barriers as well as individual actions of all attackers and defenders. Each parameter can be modified for each scenario considered to determine the projected outcome. In addition, casualty estimation and requirements for emergency response beyond the initial defensive response can be developed.
While analysis of select parameters such as response force positioning would provide little benefit for events that occur over a very short duration—the July 2012 movie theater attack in Aurora, Colo., for instance—valuable data from such models still can be gleaned to develop mitigation and longer-term response strategies to such events.
As modeling and simulation systems and cognitive task analysis science become more advanced and more integrated, detailed virtual investigations will provide an even greater role in planning effective response and recovery operations for mass casualty terrorist events. It also will benefit the subsequent development of associated training and mitigation efforts.
Such planning and preparedness training can effectively reduce the impact of active-shooter attacks at targeted facilities, both military and civilian.