Skip to content

Eliminating the Need for a 4th NYC Area Airport with Collaborative Communications Technology

Eliminating the Need for a 4th NYC Area Airport with Collaborative Communications Technology

  • Speaker: Ted Willke
  • Abstract – Presently, aircraft separation and runway usage are controlled manually, primarily through voice communications between Air Traffic Control and pilots. With existing procedures, the air system is suffering from limited runway capacity and airports, such as John F. Kennedy International in New York, and authorities are contemplating construction of an additional airport. Capacity aside, safety remains of paramount concern and yet runway incursion rates have increased steadily over the past two decades. We apply a broadcast protocol, called MRBP, to air-air communications to increase peak capacity by up to 65% for one runway and 49% for intersecting runways. The increases result from aircraft tightly coordinating their movements through shared clearances and other announcements. This shared communication also permits onboard computers to quickly detect aircraft deviations, controller errors, and communication faults. We assess the runway capacity benefits, demonstrate the protocol’s safety mechanisms, and discuss how the protocol can bring the safety benefits of controlled airspace to traffic at smaller regional airports.


  • worst airport delays in the county
    • jfk: 811/hr (100/r desired), lga: 71, 80, EWR (81 vs 100)
  • how do we fix it?
    • either new airport or new technology
    • new runways does little, better surveillance didn't help, capacity 233 to 280+ is desired
    • propose: wireless communication allowing ATC to be offloaded and decentralized - benefit all 3 NTSB challenges
  • systems today
    • networks: ground, approach, departure, tower - departure/approach has independent runways
      • enter queue for different types and handled by FIFO
    • centralized control of points impacts runway - efficiency (timeline decisions and delays), safety (miscommunication, disorientation, non-compliance)
    • responsibility: approach, handoff, spacing, sequencing, traffic routing, runway interference, compliance monitoring, instruction/clearance delivery, runway configuration
      • often don't deliver message to aircraft until right before because of other jobs

proposed system -- collaborative communications

  • delegate routine decisions to aircraft to inboard flight systems
    • cohorts maintain a shared view of traffic, requires communication supporting data consistency, with proper inputs distributed will agree on actions in fail-safe
    • several different groups
    • ATC refocuses on system management - queue count, management
  • responsibility
    • offloaded - spacing execution, runway sequencing, clearance and instruction execution, handoffs, compliance monitoring
    • system management - queue management, sequencing management, traffic routing, runway reconfiguring, exception handling

broadcast protocol (framework)

  • mobile reliable broadcast protocol - wireless medium is weakest link and requires new safeguards
    • data consistency - cohorts use same data, two aircraft will use same data in same order
      • message transmit - ACK or retry
      • if majority does not receive, throw away message wait for received - need some safe exit state
      • if majority does receive, consensus on use of data
      • aircraft can discover message was sent but doesn't have it so can go into safe state for recovery (all others want to use message)
    • deterministic reliability - messages are retried and acknowledged, sources and recipients know when they are at risk of not getting a message

token ring (protocol)

  • take a slot in the token ring and own it at a time, that aircraft is responsible for ACK'ing messages at any time
    • earliest site to ACK a message determines its location in ordered list (ACK assigns sequence number)
  • any receiver may retry an ACK or message k times
  • tokens transmitted after certain time (accounting for network delay) or retry count list the token as received or not [identify equipment failure]
    • future tokens form a vote on whether to include the ACK'd messages in the length
    • vote ends after n tokens are passed
  • timeline: message from source -> other sites ACK -> votes shared and agreed upon (vote out of group) -> survivor ACK transmitted
    • create frames of collaboration and it can change over time
    • L1 provides probabilistic reliability exceeding ADS-B
    • L2 provides new data consistency and determinism guarantees
  • example 1: aircraft message doesn't go to ATC, but cohort sends to ATC that one was missing; group accepts or reject based on voting
  • example 2: message and ACK loss due to equipment failure; aircraft either recovers with protocol or knows there is a failure because of no messages sent here

applications to capacity

  • today: interval starts when A reaches T and ends when D rolls away from T; D is cleared after wake passes and A is assured to exit, release of D from
    • runway occupancy time =  ROTA_A + d_t + d_{acq} ( d{ts} * 2 * \frac{1}{1-p} -1 ) \geq 70
  • today: departure arrival - A may arrive too early from G, ATC must ensure A does not overrun D to make sure runway is utilized, D must commit to take off before A commits
    •  capacity_{0.96} = \frac{2*3600}{t_{DA}+t_{AD}} \leq 46.3 aircraft/hour
  • what can be done:
    •  capacity_{0.96}  = \frac{2*3600}{ \epsilon + MAX(W, ROT_D) + MAX(W, ROT_A) + d_i + d_{acq, max} + 3*d_{tx, max} }
    • 70's - high speed ramps to reduce ROT, 9's model prediction of W, 00's - decrease  \epsilon due to improved ATC surveillance tech
    • remains - reduce inter aircraft arrival time, improve time line of clearances
  • arrival occupancy model - try to get about 60s for arrivals (observations show that often empty runways)
    • ROT (runway occupancy time) will fluctuate - variations in dynamics, weather, multiple exits
    • solution should provide fine grain control of capacity, reduce ATC workload, increase time for reaction


  • simplest method - aircraft following
    • ATC sets following policy, follower motion based on predecessor ADS-B, similar to automated cruise control for road vehicles
    • simulation: tell aircraft to follow, then perturb aircraft at front (for arrival time) speed changes too much
    • simulation: tell aircraft to follow with look-ahead spacing grows more; more stable sequence
  • MRBP's role in approach following
    • establishes group membership, ensures cohorts change leader in concert (synchronous), detect and deal with failures (detect failure, group partitions or falls back to safe)

clearance offload

  • today: ATC manually or automatically establishes right-of-way order
    • future: ATC gives policy (management of queue) and cohorts manage actual group operations; reduce workload
  • take-off of aircraft can send trigger to cohorts instead of waiting for ATC reaction
  • studies have shown that second "crossing" runway does not increase capacity
  • lessons - 280+ hurdle will avoid 4th airport
    • JFK - red or blue 100+/hr, LGA - (single -- 76.3/hr), EWR blue (100+/hr)

safety improvements

  • requires best of bad situation - failure cases are aircraft violates clearance or ATC sends wrong clearance
  • micro-flight plan proposed...
    • finer granularity and extend to ground operations
    • reflect performance characteristics of aircraft
    • include machine-generated extrapolation and ETA
    • include annotations of clearance status
  • consistency check
    • ATC - clearance != current, detect ATC clearance error, message enabled clearance
    • movement != current clearance or micro-FP - aircraft clearance violation or crew deviation (deviation), message uses clearances, ADS-B, micro-FP
    • revised micro-FP != ATC instructions ....
  • summary - collaborative MRBP vs. CPDLC (control-pilot data link communication)
    • makes sure it's consistent, appropriate, and properly understood
    • performance - make a timely decision, communicate without delay and reliably (already in CPDLC)
    • capacity and safety benefits of new tech are underestimated, stronger communication guarantees new opportunity, ATC distribution imp


  • assumption is that runways aren't used efficiently largely to communication delays, but when allowing aircraft to take-off right after landing, how can this be more quickly communicated?

    • w.r.t. turbulence, can use LIDAR to determine air-patterns and improve detection of weight turbulence and wind-shear; could automate it once sensors are deployed
    • sometimes you can't see the aircraft, but not counting on crew of waiting operation, counting on aircraft that is doing the action, which should know dynamically and mechanically
    • even save 10-20 seconds for each aircraft will add up over many different operators
  • what about deliberate mis-communication by rouge aircraft? (are there fail-safes from other aircraft?)

    • by mandate it's too severe to interfere with this signal, could add digital signatures and security not really addressed
  • what are similar methods deployed at other airports that are comparable? you mentioned international airports as example
    • new technology like "hold-short" lights to prevent violation or collision, new surveillance technology is getting better but still unreliable and gets lost
    • loss is that european airports allow multiple aircraft entering runway
  • talked about micro-FP to solve runway problem, would variation of TCAS help to solve this too?
    • info: TCAS (terrain/traffic collision avoidance system) - if two aircraft realize too close then can enter into conflict negotiation between peers.....
    • could, but focused on just that problem (just collision) -- could add and use TCAS but part of the problem is that it's own between two parties, but this lagged-paired reaction could have some critical error
  • pointed out that policy might help loss of ATC policy vs. aircraft trying to manage it themselves (i.e. paying attention to ABORT operations)
    • FAA changed definition only recently to define incursion as *possible* problem because of error in decision

interesting links and commentary

Post a Comment

Your email is never published nor shared. Required fields are marked *