testified in favor of a more thorough investigation of the various systems that were being proposed.

The debate was carried to the Senate Aviation Subcommittee hearings in the Fall of 1971 at which time most of those organizations and agencies testified in favor of making CAS mandatory; however, the SECANT, AVOIDS, and TimeFrequency System proponents each testified on behalf of their own concept. The FAA Deputy Administrator testified that there was insufficient information available at the time to permit selection of a specific CAS for national implementation.

By January 1972, FAA announced an "overall investigation of airborne CAS with responsibilities assigned to its Office of Systems Engineering Management." An interdepartmental Group on Collision Avoidance involving FAA, DOD and NASA was formed to promote interagency cooperation on collision avoidance and encourage joint ventures among those agencies.

In July 1972, FAA transferred funds to the Navy which, at its NADC Johnsville, let a contract with RCA to purchase RCA SECANT equipment for test and evaluation. FAA signed a contract with McDonnell Douglas in October 1972 to determine the number of Time-Frequency ground stations that would be required to provide for coverage of the U.S. In November 1972, FAA signed an interagency agreement with Navy NADC Johnsville for the procurement and flight test of the Minneapolis-Honeywell AVOIDS equipment. In late January 1973, NADC had contracted with Minneapolis-Honeywell for the purchase of the AVOIDS equipment.

United States documentation, in 1972 12 to the Seventh ICAO Air Navigation Conference, described the Time-Frequency, SECANT and AVOIDS CAS Systems, and indicated the availability of information on the FAA NAFEC firstphase evaluation of ATC/CAS System interaction in terminal areas. 13 These studies were conducted using real time simulation.

Since the relating of airborne Collision Avoidance System development that is provided herein is chronological, it is important at this time for me to digress from reporting airborne system development to note that in early 1973 a milestone was passed when FAA began evaluation of "conflict alert" at its Jacksonville Air Route Traffic Control Center. Conflict alert provides high altitude sector ATC controllers with computer-generated advance warning of potential imminent conflicts. That was the first automated backup of the separation which is expected to be provided by human controllers. Conflict alert uses the ATC Radar Beacon System data and a programming improvement to the NAS Stage A ARTCC computer. An FAA report (EM 73-7 March 1973) outlined the FAA E&D Program plans for ground-based separation assurance. Thus, it was at that time that improvement was first offered in the direction of using the automated ground ATC System as a means of detecting pilot or controller errors which would reduce separation standards below those desired.

Returning to airborne CAS System development, by November 1973 the FAA NAFEC had published its report 13 which investigated the impact on the Air Traffic Control System of the threat logic set forth by ATA Paper ANTC 117. The "landing mode" version of the ANTC 117 threat logic was evaluated.

At the RTCA Fall Assembly, Congressman Barry Goldwater, Jr. criticized FAA for its apparent opposition to an independent airborne CAS. In February 1974, FAA recommended 15 16 to the Senate that it not take steps to make airborne CAS Systems mandatory, and indicated that by no sooner than 1975 it would decide on National Standards for airborne CAS Systems. The General Accounting Office (GAO) in November 1974 17 18 stated that FAA was placing too much emphasis on ground control whereas virtually all recent midair collisions (other than airline and high-performance aircraft) have occurred when air traffic control had only one or neither aircraft under control. The GAO report urged FAA to expedite its evaluation program of airborne Collision Avoidance Systems.

In mid-1974, a progress report was made 19 on the Minneapolis-Honeywell AVOIDS System which had been delivered to NADC for flight evaluation. By late 1975, Minneapolis-Honeywell announced its AVOIDS airborne CAS was "ready for use,' ," 20 and RCA announced it was dropping its SECANT airborne CAS System.21

In early 1975, FAA's Office of Systems Engineering Management requested Mitre to begin development of an airborne CAS using the ATCRBS beacon as the cooperative element. Mitre began flight tests of the system, called BCAS (Beacon-based CAS) in mid-1975 and by October 1975 rather encouraging results were obtained.22 23 The BCAS system was initially proposed in 1970 by G. Litchford.26

Footnotes at end of attachment 2.

The attractiveness of the BCAS concept was that it would create an expenditure only for aircraft which desired to protect itself from all other aircraft that are altitude transponder equipped. To be successful BCAS must not impact ATCRBS performance, and it must perform adequately in spite of the fact that replies from many aircraft will be received simultaneously, causing synchronous garble. The Mitre BCAS concept was quite clever, it called for an interrogation rate of only two per second. The Mitre analysis, 23 (Section 4) indicates that even with substantial implementation of BCAS it would have a very small impact on the use of ATCRBS for ATC purposes.

Because the Mitre BCAS uses an omnidirectional antenna for transmitting the Mode C interrogations and receiving the replies, aircraft within 1.6 nautical miles of each other will have their replies overlapped. The Mitre analysis showed that by 1985 in the Los Angeles basin there would regularly be 3-4 targets overlapping with a possibility of as many as 40 overlapping targets. To reduce the number of overlaps, Mitre developed the "whisper/shout" technique. Whisper/ shout uses the fact that the receiver sensitivity of different transponders varies as does the antenna patterns. BCAS interrogations are transmitted at various power levels followed by a set of suppression pluses (P1 and P2) at the same power level which suppresses the transponders that have just replied so that a new interrogation at a higher power level will elicit replies from another group. The process is repeated, dividing the aircraft into multiple reply groups. Testing has shown that the whisper/shout technique is effective in dividing the responses into about four groups.

Thus far we have traced the development of three airborne Collision Avoidance Systems that require special, dedicated, cooperative airborne equipment installations to make them function. These are the McDonnell-Douglas Time Frequency System (which was endorsed and further developed by the ATA and its CAS Technical Working Group working in conjunction with a number of manufacturers), the RCA SECANT System and the Minneapolis-Honeywell AVOIDS System. One other matter remains to be reported regarding these systems-that of the approximate cost to the users for implementing these three versions of A/CAS. These costs have been developed in various ways as reported by ARINC Research 24 and FAA.25 The ACAS avionics equipment life cycle costs (which include electronics, installation, non-recurring costs and maintenance costs) vary from $633 million to $862 million, depending on which of the three systems might be selected. FAA announced at its Consultative Planning Conference on Aircraft Separation Assurance,25 that the best of the three systems (in my view, primarily a cost-oriented decision) was the Minneapolis-Honeywell FAA also stated at that Conference that since the Minneapolis-Honeywell. System woud require mandatory installation of cooperating equipment in all aircraft at a cost of $719 million, it was by no means as attractive as BCAS. Recall again, the attractiveness of BCAS is that it would utilize as the cooperating airborne system the more than 100,000 ATCRBS transponders already installed in civil aircraft, including the more than 35,000 with Mode C altitude reporting, which is necessary to make BCAS function properly.

A so-called passive form of BCAS also has received considerable attention and development effort. Beginning in the early to mid-1970's, its inventor has described,20 and others have analyzed,27 various forms of an airborne Collision Avoidance System which allow the determination of an intruder's location by listening on the two radar beacon frequencies of 1030 and 1090 MHz. This passive BCAS approach requires a change to the standard ATCRBS interrogation sequence in order to communicate the azimuth of the interrogator to the passive BCASequipped aircraft. Presently the concept calls for such azimuth marks at the eight points of the compass. The passive BCAS must identify and track the PRF of each interrogator and each responding transponder replying to its interrogation. This form of passive BCAS must also measure the time of arrival of each interrogator signal and the differential time of arrival (time intruder's reply was received and the time at which the BCAS-equipped aircraft would have heard the interrogation that elicited the reply). Passive BCAS must also compute differential azimuthwhich is the angle measured at the radar site from BCAS-equipped aircraft to the intruder. With two optimally located interrogator sites it is theoretically possible to gather sufficient information to make passive BCAS function; however, experimental versions of the concept developed by the contractor generally have required more than two ground sites. When such ground sites are not present, the passive BCAS must become active, interrogate other aircraft, and must accept and solve the same problems of any other active BCAS. Passive BCAS computes intruder position using the data just identified plus decoding of altitude and identity. Thereafter, some threat detection logic, such as Tau, would be employed and an avoiding maneuver command generated when appropriate.

Passive and active BCAS, both of which depend only on the present altitude reporting ATCRBS transponder, have absolutely no way of assuring that the escape (avoiding) maneuvers are complementary unless some air-to-air communication link is added. With passive or active BCAS, the ATCRBS altitude reporting equipped aircraft is sensed-if it is deemed to be a threat, a collision avoidance maneuver is displayed to the pilot. It would then become incumbent on the BCASequipped aircraft to avoid the other aircraft. It is obvious that this may not be either practicable or feasible in some or perhaps many instances; for example, in those instances when the BCAS-equipped aircraft does not have the capability of out-maneuvering the ATCRBS equipped aircraft, which may be already maneuvering. If both aircraft are similarly equipped with BCAS, both will be calculating an escape maneuver. It then becomes a key element in system design to ensure that these avoiding maneuvers are complementary rather than conflicting, e.g. one climb, one dive is complementary; both climb is conflicting.

To provide the proper coordination to assure complementary maneuvers, an air-to-air data link is necessary. Since this coordination occurs after a threat has been detected and the time available to provide safe separation is very short (typically 25 to 40 seconds), the integrity of the air-to-air link is extremely important. There are a number of ways of providing the air-to-air link. Attempts have been made to use the ATCRBS signal format and several of the currently unused bits of the ATCRBS reply (e.g. "X" bits of the Mode A and C replies). These proposals, and others, short of using an addressed link such as DABS, have been studied and reported on by FAA.28 Since two messages must get through, the address and the intent, the probability of successful communications for traffic densities approaching today's busy airports is low, and the probability of making an error by receiving the correct address and a false message is substantial greater than 10 percent of the time. Flight tests reported by MIT's Lincoln Laboratory show that a DABS format air-to-air link performs adequately.34

For the reasons already set forth, it has been the airline view for well over a year that for any BCAS to be succesful it must include both active interrogation and DABS capability. As of this date, most airlines believe that need for or desirability of adding any form of passive capability to BCAS remains an item that requires further development and study to permit agreed, well-informed decisions to be made.

An Airlines Electronics Engineering Committee (AEEC) sponsored Seminar on Beacon Based Separation Assurance was held as a part of the AEEC Meeting in Munich, Germany on September 3, 1976. Seven excellent papers were presented at that Seminar.29 A review of those papers should convince even the most skeptical that a lot was known about active BCAS and its related DABS air-to-air link at that time over two years ago.

The problems of the passive BCAS that were generated by its need for several ground sites and the synchronous garble problems of the active BCAS, particularly in high-density areas, led to the development of a single site (SSBCAS). It is designed to provide protection to the BCAS-equipped aircraft in the presence of only one ground site. The single site BCAS requires the installation of a modified DABS transponder at each ground-based surveillance sensor. The transponder communicates with the BCAS aircraft which is equipped to use the SSBCAS. The merit of SSBCAS is that it would permit single site operation beginning with ATCRBS transponders and continuing through full implementation of DABS. Please read Reference 30 for a detailed description of this concept.

In late 1977 FAA set up a BCAS design team led by Dr. Edward Koenke of OSEM to:

(a) Formulate and document a BCAS concept that would work in all airspace and all traffic environments.

(b) Provide some level of protection for a wide variety of users.

(c) Minimize interference with the ground/ATC surveillance and Air Traffic Control Systems.

(d) Develop an engineering requirement which defines the performance specifications for the BCAS concept.

(e) Provide a realistic work statement to fully support the concept and the engineering requirement.

The FAA report that documents the first three items has recently been made available and is in several parts, the Executive Summary (Part I) 31 and the Concept Description (Part II).32 The third volume of the concept description provides the thirteen appendices and is now being printed.

Footnotes at end of attachment 2.

The FAA BCAS design team concluded that to obtain acceptable performance of the aircraft separation assurance function in all airspace, done totally in an airborne unit, is a very complex task. A directional antenna is required in the aircraft. A family of solutions is required including active, single, and multiple site passive. The envisioned system is expected to provide measurement accuracies that are at least the equivalent to (in many cases better than) those provided by the FAA's Automated Traffic Advisory and Resolution Service (ATARS) System-based on the capability of ATARS at a 50 nautical mile range. This multifunction solution has been appropriately called FULL BCAS. The FAA BCAS design team 31 32 recommends that its BCAS be interfaced with the ATC System, even beginning with the ATCRBS environment. This would provide some control of the BCAS by the ground-based ATC System in the era before ATARS is provided by FAA. The analysis by the FAA BCAS design teams shows 31 (Table 1) that FULL BCAS would have prevented 13 of the 15 midair collisions involving air carrier aircraft in the 1964-1972 time period. The two that the FAA BCAS design team estimates would not have been prevented are those that occured over Mt. Carmel, New York and at Harlingen, Texas. The omni-directional active BCAS, FAA estimated, would have been effective for about 60% of the same collisions (9 of the 15). The FAA report also states that active BCAS performance becomes worse as traffic levels increase. For example, by 1995 it could only have been expected to prevent about 40 percent of the collisions.

One of the most important aspects of any separation assurance system, whether ground-derived, such as the FAA ATARS, or airborne, such as ACAS or the many versions of BCAS, is the threat logic. After a Collision Avoidance System gathers the data on appropriate aircraft it processes the data to determine if certain separation criteria are about to be violated. This processing takes place sufficiently in advance of the time of predicted nearest approach such that an avoiding maneuver can be displayed, the pilot will react, and the aircraft will maneuver in time to provide the desired separation. The most extensively developed, evaluated, and described airborne CAS threat logic is that of ATA Report ANTC 117.6

A computer simulation of the ANTC threat logic was undertaken 33 by McDonnell-Douglas in March of 1970. A total of 13,167 two-aircraft collision encounters were simulated using computer modeling. The report shows that for all level flight encounters, turning and non-turning, as well as climbing and diving encounters with rates up to 5000 FPM per aircraft, the separation provided by the ANTC 117 threat logic was satisfactory. Simulated pilot actions, control initiation, and execution of the maneuvers in normal and worst case flight operations were all simulated. These various conditions were loaded into the computer; the computer than "flew" the aircraft, modeled the communication, performed the evaluation as specified for the CAS equipment, modeled pilot action and reaction, and controlled the escape maneuvers of the aircraft. Appropriate printouts provide a record of the final separation of the aircraft. Work to improve the ANTC 117 threat logic began in early 1975 and has been reported in a number of papers 35 36 37 38 in work accomplished by the Institute for Defense Analysis and Mitre (METREK Division). The present work by Mitre on improving the threat logic includes uses of ARTS data tapes which provide track data on aircraft operating at typical low-, medium- and high-density areas. Mitre uses these data tapes to exercise its threat logic and determine how many alerts and maneuver commands would be generated by various versions of the threat logic. For example, recent Mitre data 39 shows that varying the Tau value between 25 seconds and 40 seconds changes the average number of alarms (pairs of aircraft) from about 5 per hour to 15 per hour at Washington National Airport with today's traffican average of 47 tracks per scan and 13 percent of the traffic VFR. Obviously much work remains to be accomplished.

Significant work on the interaction of the ANTC 117 threat logic with the ATC System has been accomplished by FAA NAFEC using its dynamic, real time simulation, and fast time computer simulation. This program, under the guidance of Gordon Jolitz, began in 1970 and has been reported in several papers. 40 41 42 43 44 One of the most important findings in these papers is that "Observation of and conversation with the controller team leads to the convincing conclusion that they, the controllers, made subtle adjustments to the techniques used in handling the arrival flow when the CAS was present in some or all of the flights."

It was also reported that under existing ATC procedures for the control of traffic during IFR conditions, virtually no ATC/CAS interaction would be found under conditions similar to those simulated by the FAA NAFEC real time simulation. What I conclude from such reports by a most competent and qualified FAA

researcher is that we must continue to refine and evaluate by all possible means, including real time dynamic simulation as well as fast time computer simulation, any necesary improvements to the BCAS threat logic so that acceptable logic that will have a minimum reaction with the ATC System can be developed in time for its use in an acceptable BCAS.

SUMMARY

Having traced the development of airborne Collision Avoidance Systems from a gleam in the airlines' eye in the mid-1950's to a complex system which is still being developed, it is my view that it is safe to say that airborne CAS has a good future. We must continually bear in mind that airborne CAS is a backup for a backup to the ATC System. The automated ground environment is the first level backup to the humans, both pilots and controllers. Currently this is provided at most altitudes by conflict alert in both enroute ATC Centers and most terminal areas equipped with ARTS III Systems. Later the DABS will provide the ability to add ATARS, which will both detect and resolve conflicts. To the degree that may be necessary, a BCAS is intended to backup that automated ground environment when it might fail or in areas of the airspace where the automated environment does not exist-in fact may never exist.

Since I have been involved in the development of airborne collision avoidance from its inception, perhaps I am in the worst position to judge fairly whether it will ever see use. On the other hand, having followed each turn in the road, and in fact plowed a few of the roads toward progress myself, I have a good sense of where we are and how far we are from a suitable solution. The very recent, most unfortunate, accident at San Diego should only serve to spur all of us to seek an early solution to the "backup to the backup" problem-which obviously has to be solved.

In my discussion today I have tried to cover those instances and events I believe to be significant. In such an involved subject I have undoubtedly missed some, but please forgive me-it was unintentional.

I am always thankful for RTCA as a forum to present problems, discuss alternative solutions, and put the agreed solutions on paper. Frankly, I think the day is close at hand when RTCA will begin to write its Minimum Operational Performance Standard (MOPS) for a suitable BCAS. May the Dear Lord give us the wisdom to accomplish the task both promptly and correctly.

FOOTNOTES

1 White, Frank C.-Is an Airborne System for Collision Avoidance Operationally and Technically Feasible? RTCA Paper 127-57/AS-183, Joint IRE/RTCA Symposium, Los Angeles, May 7-9, 1957 IRE Trans., ANE Vol. 4, No. 2, June 1957.

2 Collins Radio Company-Airborne Proximity Indicator PI-101: Descriptive Specification CDS-311, September 28, 1956.

3 American Aviation Daily Collins Drops Current PWI/CAS, Cancels Airline Orders, Vol. 106, No. 6, p. 41, Jan. 9, 1957.

Morrel, J. S.-The Use of Self-contained Range and Azimuth Measuring Apparatus to Detect Collision Courses-ION Journal; Vol. 11, No. 3, July 1958.

5 Morrel, J. S.-Fundamental Physics of the Aircraft Collision Problem, Bendix Aviation Corp., Technical Memorandum 465-1016-39, June 1. 1956.

Air Transport Association of America, Air Navigation/Traffic Control Division (ANTC) Report No. 117, Revision 2. September 29, 1967, (Att. 3 to Summary Report, Tenth Meeting ATA CAS Technical Working Group).

7 Martin-Marietta-Baltimore Division Final Report: Flight Test and Evaluation of Airborne Collision Avoidance System, Vol. 1, April 1970.

8 McDonnell-Douglas-Press Release 71-103 dated June 22, 1971.

Washington Evening Star-Piedmont is First to Sign for Safety System for Pilots (Charles Yarbrough), June 22, 1971.

10 Breckman, Jack-Avoiding Mid-Air Collisions: A Catechism of Truth or Consequences, Paper before the student chapter of IEEE at Lehigh University, October 16, 1969 by RCA Engineer, 40p.

11 Breckman, Jack-To See Or Not To See, Testimony before the National Transportation Safety Board's Hearing on Mid-Air Collisions in Washington, D.C., November 6, 1969, 7p.

12 Federal Aviation Administration, Interagency Group on International Aviation, Seventh (ICAO) Air Navigation Conference-Draft Advance Documentation (Summary Papers) for Agenda Item 7 (Systems for Collision Avoidance) for the Meeting in Montreal, 5-29 April, 1972, IGIA 72/1.135, March 20, 1972. 13 Jolitz, G.-Air Traffic Control/Collision Avoidance System Interface Simulation-Phase II, FAA, National Aviation Facilities Experimental Center, Project No. 052-241-050, Report RD73-140 (NA-73–400), November 1973, 193 p., AD 771 185, N74-12361.

14 Goldwater, B. M., Jr.-Criticism of FAA Handling of Collision Avoidance Systems, In: Upgrading the ATC System; Proceedings of the Annual Meeting, Washington, D.C., Radio Technical Commission for Aeronautics, 1973, 6 p.

15 Klass, Philip J.-Anti-Collision Systems Report Readied, Aviation Week & Space Technology, Vol. 100, No. 6, p. 38-41, February 11, 1974.

16 Electronics-FAA's Decision on Airborne CAS Delayed Till 1975, Vol. 47, No. 3, p. 59, February 7,

1974.