Mass Transit Los Angeles
Full Scale Exercise

   

USC CREATE Team

Milind Tambe
Isaac Maya
Erroll Southers
Heather Rosoff
Vern Glaser
Matthew Brown
Eric Shieh
Francesco Delle Fave
                        

Los Angeles County
Sheriff's Department Team

Lt. John Sullivan
Sgt. William Lynch
Sgt. Paul Hanley


Security Agencies Involved

Los Angeles County Sheriff's Department
  • Transit Services Bureau
  • Emergency Operations Bureau
  • Special Enforcement Bureau
USC
  • CREATE Center
  • Viterbi School of Engineering
Orange County Sheriff's Department Amtrak Police
California Public Utilities Commission Los Angeles City Fire Department
Transportation Security Administration
  • Federal Air Marshall Service
  • TSA Los Angeles
  • TSA Headquarters
Metro
  • Security
  • Rail Operations
Mozaic Solutions
IntraMetrix Consulting Inc.
Constant & Associates Inc.


Description

A Full Scale Exercise (FSE) is a specific type of real world operation designed to assess the capability of large emergency management systems in a highly stressful environment. The idea is to simulate an operation in a way that is as close as possible to a real situation. For this reason, a key objective of an FSE is the actual deployment of resources.

In the context of public transit security, the Transportation Security Administration (TSA) designed an FSE to ascertain the capabilities of the different security agencies responsible for patrolling the Los Angeles Metro System (the LA Metro) during a "high threat level" period, i.e., a period where threats to the security of the LA Metro are reported, and, consequently, where the patrolling activity of the different security agencies is increased.

This exercise took place the 16 th of May 2013. It involved 23 teams of different resources including high visibility teams, high visibility weapon teams, K-9 teams, Amtrak police, crisis response motors and threat interdiction units. These teams were deployed for one day (12 hours) to patrol 10 stations defined as High Risk Named Area of Interest (NAI 1 to 10). Figure 1 shows a map of the red line of the LA Metro system where the exercise took place.

Figure 1: The Map of the Red Line of the LA Metro





Objectives

The TSA aimed to verify two key objectives: (i) the ability of the different teams to cooperate during a "high threat level" period and (ii) the effectiveness of the patrolling activity to prevent and deter terrorist activity. To achieve this, the exercise was divided into three "sorties" of 3 hours each. Each sortie was characterized by a different resource deployment and was defined as follows:

Sortie 1:
The deployment schedule was defined manually by the Los Angeles Sheriff¡¯s Department (LASD). The key objective was to achieve a uniform, cumulative, 25% coverage of each NAI, i.e., 45 minutes (out of the total time of 3hours) for each station (450 minutes in total).

Sortie 2:
The deployment schedule was randomized using a game theoretic approach defined by the USC CREATE team. Each team was provided with a mobile application used to visualize the team¡¯s specific schedule. The key objective was to achieve a weighted 50% cumulative coverage of each NAI, i.e., 900 minutes of coverage divided between the different stations, such that the most important stations were covered (i) more frequently and (ii) for a longer period of time than the other stations.

Sortie 3:
The deployment schedule was randomized using the same game theoretic approach the same mobile application used in sortie 2. The key objective, however, was to achieve a weighted 25% cumulative coverage of each NAI, i.e., 450 minutes of coverage divided between the different stations, such that the most important stations were covered (i) more frequently and (ii) for a longer period of time than the other stations.

The behavior of each team during each sortie was monitored by a mobile command center (see Figure 2). During each sortie different evaluators, positioned within each of the 10 NAIs, were observing the behavior of the different teams to verify the degree to which each of the two objectives defined above was satisfied (see Section Evaluation for more details).

            

Figure 2:The Mobile Command Center

            

Figure 3: Inside the Mobile Command Center




Research Challenges

The key research challenges for the USC CREATE team were to develop the game theoretic approach, the mobile application and monitor their deployment during the FSE.

The game theoretic approach was inspired by recent work on coordinating joint activities in security games. The key idea was to encapsulate the capability to cooperate and the abilities of the different teams deployed during the FSE within a security game model, which was then solved using a branch and price approach. The mobile application was built upon a similar application used for the fare evasion problem (see the Figure below).

Figure 4: The Mobile Application

Before the sorties, each team was instructed on how to use the application and read the schedule (See Figures 5 and 6). Each evaluator (Figure 7) was given a questionnaire about the level of threat / security that he was observing in the NAI where he was positioned during each sortie.

            
            
            
Figure 5:Team training Figure 6: Team training 2


Figure 7: Evaluators

Currently, the data collected is still being evaluated. Some initial results, depicted in Figure 8 shows that the majority of the evaluators felt more secure during the sortie using the 25% "game theoretic" deployment rather than the manual one.

Figure 8: Initial results