USE OF A PROTOTYPE AIRBORNE SEPARATION ASSURANCE SYSTEM FOR RESOLVING NEAR-TERM CONFLICTS DURING AUTONOMOUS AIRCRAFT OPERATIONS Richard Barhydt* Dr. Todd M. Eischeid† Michael T. Palmer‡ NASA Langley Research Center Booz-Allen & Hamilton David J. Wing* Hampton, VA 23681 Hampton, VA 23681 NASA Langley Research Center Hampton, VA 23681 ABSTRACT evaluate these design changes and consider integration issues between ASAS and existing Airborne Collision Avoidance Systems (ACAS). NASA is currently investigating a new concept of operations for the National Airspace System, designed to improve capacity while maintaining or improving INTRODUCTION current levels of safety. This concept, known as Distributed Air/ Ground Traffic Management (DAG- NASA is investigating a new concept of operations for TM), allows appropriately equipped “autonomous” the future National Airspace System (NAS). This aircraft to maneuver freely for flight optimization while concept, known as Distributed Air/Ground Traffic resolving conflicts with other traffic and staying out of Management (DAG-TM), enables pilots flying special use airspace and hazardous weather. In order to appropriately equipped “autonomous” aircraft to have perform these tasks, pilots use prototype conflict more flexibility in choosing optimal flight trajectories detection, prevention, and resolution tools, collectively while simultaneously resolving conflicts with other known as an Airborne Separation Assurance System aircraft, special use airspace, and hazardous weather.1 Air (ASAS). While ASAS would normally allow pilots to Traffic Service Providers still issue traffic flow resolve conflicts before they become hazardous, management constraints to autonomous aircraft and evaluation of system performance in sudden, near-term provide traffic separation services to those aircraft unable conflicts is needed in order to determine concept to participate in autonomous operations. A critical feasibility. component to the feasibility of DAG-TM operations is the effectiveness of an Airborne Separation Assurance An experiment was conducted in NASA Langley’s Air System (ASAS) aboard the autonomous aircraft. Traffic Operations Lab to evaluate the prototype ASAS for enabling pilots to resolve near-term conflicts and Separation assurance systems would normally enable examine possible operational effects associated with the pilots operating in a DAG-TM environment to respond to use of lower separation minimums. Sixteen traffic conflict situations well before the conflicting commercial airline pilots flew a total of 32 traffic aircraft poses a hazard to safe flight. Previous studies scenarios that required them to use prototype ASAS have shown that pilots are able to use prototype ASAS tools to resolve close range “pop-up” conflicts. tools effectively to resolve longer-term conflicts.2-3 Required separation standards were set at either 3 or 5 NM lateral spacing, with 1000 ft vertical separation One can envision situations that may arise, however, that being used for both cases. Reducing the lateral do not allow this normal opportunity for strategic separation from 5 to 3 NM did not appear to increase decision-making. Various non-normal events could operational risk, as indicated by the proximity to the require the autonomous aircraft pilot to use the ASAS to intruder aircraft. Pilots performed better when they resolve a near-term conflict. Examples may include an followed tactical guidance cues provided by ASAS than aircraft descending rapidly due to an emergency or an when they didn’t follow the guidance. In an effort to improper maneuver by an aircraft responding to another improve compliance rate, ASAS design changes are conflict. Demonstration of a pilot’s ability to effectively currently under consideration. Further studies will use ASAS to resolve near-term conflicts and regain lost *Aerospace Engineer † Senior Consultant ‡ Senior Member, AIAA, Aerospace Engineer 1 American Institute of Aeronautics and Astronautics separation is needed in order to assess concept the resolution) as a potential safety risk.9 If near-term feasibility. conflicts with lower separation minimums were determined to be more challenging for pilots, the severity These types of non-normal events will require the of these risks could be even greater. ASAS to provide effective alerts and resolutions prior to the time that an Airborne Collision Avoidance An experiment was conducted in NASA Langley’s Air System (ACAS) would give a Resolution Advisory Traffic Operations Lab to address issues related to ASAS (RA). When an RA is issued, a pilot must take use for near-term conflicts with a potential reduction in immediate action in order to avoid a potential near miss separation minimums. The experiment had the following or collision. The Traffic Alert and Collision Avoidance two primary objectives: System (TCAS) II currently functions as an ACAS aboard commercial aircraft. Depending on the own • Evaluate the effectiveness of prototype aircraft’s altitude, RA’s are only issued 15-35 seconds ASAS tools in enabling the pilot to safely prior to the Closest Point of Approach (CPA).4 Prior to resolve near-term conflicts. an RA, DAG-TM pilots operating autonomous aircraft must rely solely on ASAS for resolution guidance. • Compare the effect of 3 and 5 NM lateral separation standards (with 1000 ft vertical An additional area of DAG-TM concept feasibility separation) on a pilot’s ability to safely relates to a potential reduction in separation standards. resolve near-term traffic conflicts. Lower separation standards will likely improve NAS efficiency and capacity.5-6 Current separation EXPERIMENTAL APPROACH minimums are based in large part on the capabilities of older radar systems.7 Safety assessments are needed to Air Traffic Operations Lab determine the feasibility of reduced separation minimums. They will give strong consideration to The Air Traffic Operations Lab at NASA Langley surveillance system performance, including accuracy, Research Center is a medium fidelity PC workstation- integrity, and availability.8-10 Candidate surveillance based facility, enabling simultaneous operation by up to 8 systems include Automatic Dependent Surveillance- subject pilots. Each pilot station consists of a transport Broadcast (ADS-B) and multi-lateration systems. aircraft model and flight deck displays designed to Considering studies done for Reduced Vertical replicate the MD-11. Figure 1 shows the display and Separation Minimums (RVSM) operations, it is likely flight control suite made available to the pilots during the that flight technical errors will also be considered.11-12 experiment. Traffic information was superimposed on the navigation display and pilots were provided with a In addition to a thorough evaluation of surveillance Navigation Display Control Panel (NDCP) to adjust the system performance, a potential decision to lower the display’s traffic-specific features. The NDCP was located separation standards should also take operational to the left of the Glareshield Control Panel. All control of considerations into account. An ASAS Safety the simulation aircraft was done through the Glareshield Assessment study identified improper maneuvering in Control Panel and Multifunction Control Display Unit response to a conflict (due to ambiguous or improper (MCDU) associated with the Flight Management System resolution commands or a pilot’s failure to comply with (FMS). No manual flight control was available. 2 American Institute of Aeronautics and Astronautics Figure 1. Pilot Displays and Control Panels Subjects Scenario Set-up and Pilot Tasks Sixteen commercial airline pilots with experience in All pilots were flying autonomous aircraft in a DAG- Airbus glass cockpit or MD-11 aircraft participated in TM en-route environment and were asked to maintain the experiment. All were active pilots or had retired standard separation from other air traffic and from within the previous year. Their ages ranged from 32 to restricted airspace. Pilots were allowed to maneuver 57 and flight experience ranged from 4,200 to 23,000 freely without contacting a controller. This experiment hours. focused only on air-air separation assurance involving “autonomous” aircraft and the simulation did not Design incorporate a ground component. Future studies are planned that will look at integrated air-ground The experiment used a single-factor within-subjects applications involving aircraft of mixed equipage. design. This factor was the required lateral separation from other aircraft and was set at either 3 or 5 NM. The In addition to maintaining traffic separation, pilots were required vertical separation was 1000 ft for both cases. given a downstream waypoint to cross at a Required Time of Arrival (RTA). En route to the RTA, the Pilots flew one scenario for each lateral separation aircraft’s programmed flight path went through a 65 condition and each scenario had one designed near-term NM wide corridor with restricted airspace areas on each conflict. Although other background aircraft were side. Pilots were asked to continue to the RTA present, the data analysis only considered the designed waypoint when able after resolving a conflict. The conflict. Scenarios for the lateral separation condition designed conflict occurred about 15 minutes into a 25- were counterbalanced for order and were included as minute scenario. part of a larger study that consisted of nine scenarios flown per pilot. Several of the other seven scenarios In order to simulate an unexpected near-term conflict, were run in between those for this experiment, in order the designated intruder aircraft was hidden from the to help prevent the pilot from anticipating the designed subject pilot until just before the predicted loss of conflict. This paper only discusses results from the two separation. At this point, a “pseudo pilot” would turn near-term conflict scenarios. the intruder toward the subject and turn on the intruder’s ADS-B broadcast. It appeared about 6 NM away from and at co-altitude with ownship. An alert occurred shortly thereafter. 3 American Institute of Aeronautics and Astronautics The initial conflict geometry (proximity, approach angle, was based on the concept described by RTCA’s and time to closest approach) was designed to be the Airborne Conflict Management (ACM) committee.10 same for both 3 NM and 5 NM separation zone cases. The separation zone represents the minimum legal Pseudo pilots were used to ensure the conflict occurred, separation around an aircraft. It had a 3 or 5 NM even if the subject pilot had deviated from the original radius, depending on the scenario. The vertical programmed flight plan. Use of the same initial conflict dimension was 1000 ft for both cases. A conflict was conditions for all scenarios allowed a comparison of risk defined as a predicted loss of separation and a incurred between the 3 NM and 5 NM separation zone separation violation occurred if an aircraft penetrated conditions. another aircraft’s separation zone. The collision zone’s radius and height above/below were 900 ft and 300 ft, For the 5 NM separation zone condition, separation loss respectively. Penetration of the collision zone was occurred earlier, but further away from the intruder when considered to be a near miss. compared to a 3 NM zone. These differences would likely highlight any variation in pilot performance due to Separation Zone an aircraft being inside or outside the separation zone. As discussed below, alert symbology changed slightly 1000 ft when separation was lost. Because the intruder in the conflict scenarios always 300 ft appeared close to the ownship, a highly aggressive maneuver would have been required to avoid a separation violation. In order to evaluate a pilot’s ability 900 ft to regain separation, avoiding a separation loss was Collision Zone designed to be nearly impossible for the scenarios used in this experiment. 3 or 5 NM Depending on selected map range, the subject pilot could see additional scripted aircraft on the traffic display Not to Scale while en-route to the RTA waypoint. An automatic traffic filter showed other aircraft determined to be Figure 2. Separation and Collision Zones operationally significant, based on proximity, current trajectory, and intent. Figure 3 shows a notional timeline for ASAS and ACAS as applied to this experiment. After a conflict After each scenario, pilots were asked to complete a was detected but prior to separation loss, ASAS questionnaire asking them to rate the effectiveness of the provided information to assist the pilot in maintaining traffic alerts and resolution guidance in resolving the adequate separation. Strategic and/or tactical conflicts. They were also asked to assess the safety risk separation assurance decision support tools were posed by the pop-up conflicts. available, depending on the time to conflict. The strategic resolution system was integrated with the FMS Displays and Alerts to provide an efficient conflict resolution and return to the programmed flight plan. Due to the short timeframe The prototype ASAS combined a state-based (position, of the planned conflicts, pilots were not able to use ground speed, and ground track) tactical system with a strategic resolutions for this study. Tactical conflict strategic system based on state and FMS intent resolutions used only state information and provided information. ASAS tools were part of larger decision recommended changes in heading and/or vertical speed. aiding system, known as the Autonomous Operations Tactical resolutions became available 5 minutes prior to Planner.13 TCAS was not available for the experiment. the conflict and strategic resolutions were phased out 2 minutes prior to conflict. This transition region was Traffic alerts were based on outer and inner zones outside the timeframe of the pop-up traffic. Further surrounding the aircraft, referred to as the separation and details of these systems are provided by Wing et al.3,14 collision zones, respectively (see Figure 2). This design 4 American Institute of Aeronautics and Astronautics If separation was lost, tactical resolutions continued to The experiment focused on maneuvers performed by provide guidance to the pilot on ways to regain the pilots just prior to when TCAS would have provided an minimum separation. In these cases, pilots were RA. As discussed earlier, TCAS issues RA’s between instructed in a pre-flight briefing to follow the tactical 15 and 35 seconds prior to CPA. Conflict alerts during guidance. If applicable, an alert was issued one minute the experiment occurred mainly between 60-70 seconds prior to predicted collision zone entry (considered on the prior to the CPA. This timeframe may be considered to timeline to be close to the CPA). Note that the loss of be near a transition between separation assurance and separation could shift left or right and occur before or collision avoidance. Because TCAS was not after a collision zone alert, depending on the relative incorporated, no explicit effort was made to consider velocity between the two aircraft. For this experiment, a integration of ACAS RA’s and ASAS resolutions in collision zone alert, if issued, normally occurred prior to this simulation. separation loss. Issuance of this alert did not affect the tactical resolutions. Figure 4 shows Primary Flight and Navigation/Traffic Displays as typically seen by the pilot shortly after the intruder aircraft appeared. The intruder was well inside Collision Zone the conflict detection threshold and therefore the Alert CPA conflict alert was generated almost immediately. This alert was shown as an amber band along the ownship’s Strategic Tactical TCAS flight path where the separation loss was predicted to Separation Separation RA occur. The intruder was also color-coded amber to Assurance Assurance Separation represent its threat level. Lost When a state-based conflict was detected, pilots were presented with a tactical resolution, shown as green bugs on the vertical speed and heading indicators. The 2-5 1 0 ~0.5 0 bugs showed the recommended change in either heading or vertical speed needed to resolve the conflict. This resolution guidance allowed the pilot to resolve the conflict by following either the lateral or vertical Time to C onflict (min) guidance alone. The resolution algorithm used a modified voltage potential method, originally developed by Eby15 and refined by the NLR16. This Time to algorithm acts to increase the projected separation CPA (min) between the two aircraft at CPA. Experiment’s Area of Concentration Conflict prevention bands were placed on the heading and vertical speed indicators to show headings and Prior to Separation Loss vertical speeds that, if flown, would cause a conflict Time to Conflict (min) with another aircraft. These bands were predicted using After Separation Loss current state information for up to a five-minute time horizon. The heading and vertical speed bands Collision Avoidance Region assumed a constant ground speed maneuver in either Time to CPA (min) the horizontal or vertical plane, respectively. A change Notes: Times Not to Scale in speed or a combined lateral/vertical maneuver would TCAS Not Available for Experiment change the conflict prevention bands. Once inside another aircraft’s separation zone, conflict prevention bands were removed. After separation loss, the bands Figure 3. Notional ASAS/ACAS Timeline no longer provided useful information because they would span the entire heading and vertical speed range. 5 American Institute of Aeronautics and Astronautics Conflict Restricted Airspace Prevention Band Tactical Resolution Loss of Intruder Separation Band Ownship Figure 4. Primary Flight (left) and Navigation (right) Displays Results In order to compare results from the 3 and 5 NM scenarios, all results using ε were calculated with a 5 NM separation zone. With this convention, the same Performance Metrics value of ε corresponded to the same threat level for both separation standards. For cases where the tested A single parameter that combines the relative lateral and separation zone was 3 NM, ε<1 did not necessarily vertical distances between the ownship and intruder imply a separation violation (Figure 5, Region A). aircraft was used as a measure of threat severity. This Conversely, a separation violation may still have parameter, referred to as epsilon (ε), is discussed by den Braven17. It represents the ellipse distance of an ellipsoid occurred when ε>1, for cases where the aircraft was outside the ellipsoid, but within the separation zone enclosed within the cylindrical separation zone. cylinder (Figure 5, Region B). Figure 5 shows a cross-section of this geometry for an Two performance metrics were considered for the ellipsoid corresponding to a 5 NM separation standard. experiment: The lengths of the ellipsoid’s major and minor axes are the separation zone’s diameter and total height, • Threat Proximity (ε ): actual minimum respectively. When ε=1, an aircraft is on the surface of min ε between the two aircraft. the ellipsoid centered on another aircraft. Once inside the separation zone, threat severity increases as ε approaches • Risk Mitigation (ε ): difference zero. diff ε = 1 between the predicted minimum ε at the time the alert was issued (based on current state information for both ε = 0 1000 ft aircraft) and ε . min Threat Proximity 3 NM Figure 6 shows εmin for the 3 and 5 NM separation cases, broken down by whether the pilot complied or 5 NM did not comply with the tactical resolution. The pilot A was said to comply if his initial maneuver (either B lateral or vertical) was in the same direction as the corresponding tactical resolution. Using this Figure 5. Ellipsoid and Cylindrical Separation Zone definition, the pilot complied 9/15 (60%) times for 6 American Institute of Aeronautics and Astronautics the 3 NM separation zone and 13/16 (81%) times for the 5 NM/comply cases, with an overall mean across all NM zone. Note that one 3 NM separation scenario was scenarios of 0.04. lost due to an earlier pilot maneuver that prevented the planned conflict from taking place. Figure 7 shows ε as a function of separation zone diff size and resolution compliance. A linear regression of tactical guidance compliance (combined across Complied with tactical guidance 1.0 separation zone conditions) onto ε showed that diff Did not comply with tactical guidance compliance could marginally predict ε (F(1, 29) = 0.8 diff 3.118, p = 0.088, R2 = .066) and that ε was n min0.6 marginally larger forad jthose who complieddiff with o sil 0.4 guidance (M = .446, SD = .297, n = 22) than those p E who did not comply (M = .256, SD = .192, n = 9). 0.2 Observed power (1 - β) was .400. This result was not 0.0 significant at the α = .05 level used for the present n=9 n=6 n=13 n=3 3 NM Separation 5 NM Separation study. Note: Error bars represent 1 standard error of the mean. The lateral separation zone distance did not appear to Figure 6. Minimum Epsilon by Separation Zone Size and affect the pilot’s risk mitigation. No significant Tactical Resolution Compliance differences were found for ε between the 3 and 5 diff NM separation cases when combined over pilot The lateral separation zone distance did not appear to compliance. affect threat proximity. No significant differences were found for ε between the 3 and 5 NM separation cases min when combined over pilot compliance. 1.0 Complied with tactical guidance 0.9 Did not comply with tactical guidance A linear regression of tactical guidance compliance 0.8 (combined across separation zone conditions) onto εmin n diff 00..67 was performed. The regression showed that compliance o 0.5 could significantly predict ε (F(1, 29) = 4.264, p < .05, sil 0.4 min p R2 = .098) and that ε was larger for those who E 0.3 adj min 0.2 complied (M = .497, SD = .310, n = 22) than for those 0.1 who did not comply (M = .266, SD = .193, n = 9). 0.0 n=9 n=6 n=13n=3 Observed power (1 - β) was .515. 3 NM Separation 5 NM Separation Note: Error bars represent 1 standard error of the mean. Risk Mitigation Due to Maneuver Figure 7. Differences between Actual Minimum Epsilon and Minimum Epsilon Predicted when Alert Whereas ε is a measure of maximum threat severity, min Issued, by Separation Zone Size and Tactical ε can be used to assess a pilot’s ability to improve the diff Resolution Compliance conflict situation over that predicted when the alert was first issued. Epsilon is a more suitable metric for diff Pilot Questionnaires comparing pilot performance because it accounts for any differences in predicted ε that may have occurred due min Post-scenario questionnaire results related to the to different initial conflict geometry. Although the decision support tools and the pilots’ perceived experiment was designed to generate the same initial operational safety are shown in Table 1. Each conditions (proximity, approach angle and time to closest question had a 1 to 7 rating scale, ranging from 1 point of approach) across all scenarios, prior maneuvers (least favorable) to 7 (most favorable). The scale by the subject pilot caused differences in a few cases. On description as applied to each question is given in the average, however, these differences were minimal. 2nd column of Table 1. Predicted ε at the time of the alert ranged from 0.01 for min the 3 NM/did-not-comply scenarios to 0.07 for the 5 7 American Institute of Aeronautics and Astronautics Table 1. Pilot Questionnaire Results subsequent maneuver times would lead to lower risk in cases where a conflict situation exists prior to the Question Rating Scale Overall alerting time horizon. Because a 3 NM separation Mean zone represents a smaller target, trajectory How intuitive 1: not at all 5.0 uncertainties may cause a shorter alert time when was the conflict intuitive → 7: compared to a 5 NM zone. A smaller separation zone alerting system? very intuitive may also reduce the number of alerted conflicts that How acceptable 1: not at all 4.3 evolve into safety critical situations. Due to the were the tactical acceptable → 7: consistent alert time associated with the pop-up resolutions? completely acceptable conflicts in this experiment, no such changes could What was the 1: completely 4.0 be observed. As discussed earlier, extensive analysis level of safety for unsafe → 7: is also needed to determine whether surveillance this scenario? completely safe system performance can support lower separation How did the 1: greatly 4.8 minimums. conflict manage- increased risk → ment tools affect 7: greatly Pilots were able to reduce threat severity and mitigate the risk level? decreased risk risk in response to pop-up conflicts more effectively when they followed the tactical guidance from ASAS. In order to further improve upon these Results suggest that pilots had a weak acceptance of the results, several design enhancements are currently conflict alerts and resolutions. These results may be under consideration for the conflict detection, partially due to pilots only having an opportunity to use prevention, and resolution systems. These changes the ASAS tools during high-risk short-term conflicts. focus on a more integrated approach for showing Potential design changes to the ASAS tools currently information to pilots. Previous research suggests that under consideration and discussed below may also when provided with alerts that are consistent with improve these ratings for future studies. It is not displayed information, pilots are more likely to surprising that pilots felt neutral about the level of safety comply promptly with resolution guidance.18 corresponding to a pop-up conflict. These conflicts were designed to pose a safety hazard associated with a non- As discussed above, the prototype ASAS employed normal event. during this experiment consisted of tactical and strategic separation assurance. Tactical information The DAG-TM concept incorporates several safety included state-based conflict prevention and tactical measures to reduce the likelihood of the type of scenario resolutions. The strategic system detected conflicts evaluated in this experiment1. Before maneuvering, pilots using a combination of state and FMS-based intent of autonomous aircraft must ensure that they do not create information and provided FMS resolutions. These near-term conflicts. Refined ASAS tools relying on logic differences may have caused the tactical and highly dependable surveillance systems should provide strategic systems to display different alert levels for a these aircraft with adequate warning of upcoming threat aircraft. conflicts in all nominal cases. During normal operations, pilots of both aircraft will have the necessary tools to Although pilots did not observe the transition from resolve conflicts. New decision support tools are also strategic to tactical separation assurance for the short- expected to further assist air traffic service providers in term conflicts studied in this experiment, they did fly providing separation services for aircraft not equipped for several scenarios with long-term conflicts as part of a autonomous operations. larger study.19-20 If pilots encountered cases of non- supporting (dissonant) alerts between the strategic Discussion and ASAS Enhancements and tactical systems, their overall confidence in the ASAS may have been adversely affected.21 Reducing the lateral separation from 5 to 3 NM did not appear to increase operational risk for the near-term Initial design enhancements to the prototype ASAS conflicts studied in this experiment. Additional studies have focused on the interaction between the state- could consider whether any reduction in alert and 8 American Institute of Aeronautics and Astronautics based Conflict Prevention (CP) system and the Conflict achieve a more unified mental model of conflict Detection (CD) system that considers both state and FMS situations. intent. The latter produces a Loss of Separation (LoS) band along the ownship flight path where a separation The left pane of Figure 8 shows a “false negative” violation is predicted to occur. Knowledge of intent situation where the existing CP system does not information can cause the LoS band to reflect a different validate a detected intent-based conflict. The LoS level of alert than the state-based CP band on the map band indicates that a traffic aircraft is in conflict with display’s compass rose. The CP band shows a range of ownship, however, no band is visible on the heading headings that, if flown, would cause a conflict with arc. Since a conflict exists, a CP band should another aircraft. Incorporating FMS intent information encompass the present heading, as shown in the re- into the CP system will enable the CD and CP systems to designed display (right pane of Figure 8). generate consistent alerts, thereby enabling pilots to Figure 8. “False Negative” Situation Occurring with Current Display (left) and Re-designed Display Incorporating Intent-based CP Logic (right). Conversely, the “false positive” is another situation that a band over the current heading. This indication may may cause a difference in conflict indications. The left suggest a maneuver when none is required. In pane of Figure 9 illustrates how such a scenario could contrast, the CD system has knowledge of intent develop with the existing display. The traffic aircraft information and does not detect a conflict. The on the right has a flight path that does not place it in enhanced CP system shown in the right pane of Figure conflict with ownship. Using the state projection, the 9 considers traffic aircraft intent information. CP system believes that a conflict does exist and draws 9 American Institute of Aeronautics and Astronautics Figure 9. “False Positive” Situation Occurring with Current Display (left) and Re-designed Display Incorporating Intent-based CP Logic (right). In addition to the integration of CD and CP systems, may be further increased by design enhancements to the design enhancements to the tactical resolution system prototype ASAS. are also under consideration. These changes could tap into the extensive development and evaluation of Before air-air separation assurance or other applications TCAS. As the time to closest point of approach with lower separation minimums are implemented, a decreases, it becomes more important for the ASAS thorough Operational Safety Assessment will need to resolutions to be compatible with RA’s that are be performed. This process has been described by eventually provided by ACAS.8 TCAS design goals Zeitlin.8 Included in this process is an Operational that could be considered for ASAS enhancement Hazard Analysis where potential hazards, the include: attempt to avoid crossing the intruder’s operational effects of those hazards, and mitigation altitude (especially for very close encounters), provide a strategies are assessed. Evaluation of pilot performance resolution that does not require the ownship to change using ASAS tools will likely be an important part of direction if currently maneuvering, allow time for each this process. aircraft to initiate a maneuver prior to reversing the direction of an RA, and only reverse an RA if needed to References ensure safety.4,22-23 These design goals are likely extensible to both lateral and vertical maneuvers. For 1NASA, “Concept Definition for Distributed example, previous studies have shown that pilots are Air/Ground Traffic Management (DAG-TM),” Version less likely to turn toward the intruder as they get closer 1.0., 1999. to the conflict.24 2Cashion, P. and Lozito, S., “The Effects of Different Conclusions Levels of Intent Information on Pilot Self Separation Performance,” Proceedings of the 10th International For near-term conflicts, a pilot’s ability to reduce threat Symposium on Aviation Psychology, Vol. 1, Ohio State proximity and mitigate risk appears to depend more on University, Columbus, OH, 1999, pp. 177-182. compliance with the ASAS resolution guidance than on the size of the lateral separation zone. Compliance with 3Wing, D., Barmore, B., and Krishnamurthy, K., the tactical resolutions led to improvement in threat “Airborne Use of Traffic Intent Information in a proximity and risk mitigation. These positive effects Distributed Air-Ground Traffic Management Concept: 10 American Institute of Aeronautics and Astronautics