How Do We Measure Safety Margins?
The Aviation Safety Monitor measures safety margins by estimating the frequency, duration, and severity of buffer encroachments. A buffer encroachment occurs when an aircraft pair is predicted to be closer than the minimum separation standard – both lateral and vertical – for that airspace. The encroachment algorithms account for aircraft type (heavy and other aircraft), airspace rules, horizontal distance between successive aircraft, and vertical separation. The FAA has defined minimum separation standards when aircraft are under Instrument Flight Rules (IFR). In this research, the buffer encroachment algorithms use the separation standards as a baseline to define the adequate safety margin levels.
The buffer encroachment algorithms do not identify loss of separation events. Instead, they evaluate what would occur if aircraft continued on their current speed and heading. Avoidance of a loss of separation when a buffer encroachment occurs requires action by flight crews and controllers. In many cases, these actions call for the aircraft to follow well-defined procedures accurately and in a timely fashion. In that sense, when the algorithm identifies a buffer encroachment it highlights a situation in which pilots and controllers must operate with a shared awareness of what all aircraft in the airspace will be doing. It is with this understanding that we interpret buffer encroachments as a measure of safety margins in the airspace.
The buffer encroachment algorithms use flight track messages from the FAA’s System Wide Information Management (SWIM) data service. The track data from SWIM consists of asynchronous position reports for all aircraft in the terminal airspace. To calculate the buffer encroachments, the actual aircraft trajectory must be linearly interpolated between any two given position reports. For a given pair of aircraft with piece-wise linear trajectories, a buffer encroachment can occurif any pair of line-segments that constitute the trajectories violate the buffers. The algorithms identify both the occurrence of an encroachment and its duration. The encroachment duration calculation is applied to all pairs of line segments for all aircraft pairs (see Figure 1). The sum of all segment-wise durations gives the total encroachment duration for the airspace. The Aviation Safety Monitor calculates updates of the buffer encroachments and other hazards and metrics every minute. The results are output in five-minute epochs, or time periods.
Figure 1. Example piece-wise linear trajectories for an aircraft pair
The algorithms that identify buffer encroachments allow separation thresholds to vary under different meteorological conditions. The standard separation thresholds are most relevant for instrument approaches or when aircraft are under direct controller guidance. When cleared for visual approaches under appropriate meteorological conditions, aircraft may come closer than the separation standards. This is one reason why the number and duration of encroachments increases during VMC and it is an open question what this implies for aviation safety. As the aircraft are closer, on that dimension safety margins are definitely smaller, but this may be offset by the enhanced situation awareness provided by the pilot’s visual identification of other aircraft nearby.
For the Aviation Safety Monitor, we identify encroachment events under different meteorological conditions according to the Separation Index methodology described by the FAA in its Air Traffic Quality Assurance document [1]. The Separation Index formula, defined in Figure 2, classifies Separation Conformance into four categories that measure the degree of deviation from the separation standard. Separation Conformance is measured by the index S: a value of 1.0 or higher indicates conformance with the standards. A value below 1.0 shows non-conformance, with the degree of non-conformance classified into four categories: Proximity Event, Class C, Class B, and Class A as defined in the table reported in Figure 2.
Figure 2. Separation Conformance Using the Separation Index
In the Aviation Safety Monitor, an encroachment is counted under VMC when aircraft pairs have an S value of 0.75 or below. This threshold value is based on research performed by NASA under the TSAFE project [2]. For Marginal Visual Meteorological Conditions (MVMC), an S value under 0.90 defines a buffer encroachment. Under all other meteorological conditions, an encroachment occurs when the separation index S is below 1.0. The monitoring system tracks the Separation Conformance category of each buffer encroachment using the FAA categories as reported in Figure 2. All metrics including buffer encroachments are calculated every minute and reported for the past five minutes. For example, at 06:35:00, all events and metrics are measured for the five minutes from 06:30:00 through 06:34:59. The five-minute time period used to count hazards and metrics allows for continuous updating of hazards, while allowing for a convenient and computationally efficient interval for calculating moving averages and other metrics. At 06:36:00, the calculation processes repeat and update all output files.
The 26 airports tracked in the Aviation Safety Monitor are combined into 17 metroplexes when reporting safety margins. Hazard exposure and non-separation risk metrics are tracked and reported for all 26 airports. Many of these terminal airspaces have more than one commercial airport inside the 50-mile radius. To ensure that encroachment events are counted accurately, all aircraft within the metroplex are analyzed simultaneously and reported together. For example, in the Dallas-Fort Worth area, buffer encroachments for DFW and DAL are analyzed jointly and reported out as a single metroplex. Chicago O’Hare and Midway Airport are similarly reported together for the Chicago metroplex.
References
-
Federal Aviation Administration, “Air Traffic Quality Assurance,” Order JO 7210.56C, Washington, D.C., July 2009.
-
Huabin Tang, “Tactical Separation and Safety-Alerting System for Terminal Airspace,” Journal of Air Transportation, Vol. 28, October-December 2020, pp. 184-194.
