3.2.4.1 When the pilot of an ACAS aircraft receives an RA and manoeuvres as advised, the ACAS aircraft will normally be able to avoid the intruding aircraft provided the intruder does not accelerate or manoeuvre so as to defeat the RA response of the ACAS aircraft.
3.2.4.2 If the intruding aircraft is equipped with ACAS, a coordination procedure is performed via the air-to-air Mode S data link to ensure that the ACAS RAs are compatible
3.17.3 Reduction in the risk of collision
3.17.3.1 Status of the logic risk ratio
3.17.3.1.1 The risk ratio calculated for the purposes of Annex 10, Volume IV, Chapter 4 is a measure of the performance of the logic and not the ACAS as a whole. For example, ACAS can prevent a collision by prompting the pilot to carry out a successful visual search for the intruder and it can fail because a track is not established or the pilot ignores the RA; these are aspects of the total system that are not reflected in the calculations required for Annex 10, Volume IV, Chapter 4.
3.17.3.1.2 When considering the relevance of the “logic risk ratio” figures calculated for Annex 10, Volume IV, Chapter 4 to operations or policy decisions, it might be helpful to regard them as solely the reliability that can be attached to RAs. They express the effect that following an RA will have on the immediate risk of collision when, at the time it is issued, the pilot has no information other than the RA on which to base a decision whether to follow the RA or ignore it. As a rough guide, the collision risk created by ACAS arises from following the RA so the logic risk ratio overstates this “induced risk ratio”; on the other hand, it also overstates the capability of ACAS to prevent collisions because of the many other failure modes in the total system.
3.17.3.3 Induced and unresolved risk
3.17.3.3.1 It is not sufficient to demonstrate that ACAS will prevent collisions that might occur in its absence. The risk that ACAS logic could cause collisions in otherwise safe circumstances must be fully considered, not least because in managed airspace the number of encounters potentially facing an induced risk greatly exceeds the number of near collisions.
3.17.4 Compatibility with ATM
3.17.4.1 Nuisance alert rate
3.17.4.1.1 ACAS is required to diagnose a risk of imminent collision on the basis of incomplete information. Furthermore, this information has to be independent of that providing the primary basis for aircraft separation. It follows that there will be alerts in encounters where, from an operational perspective, there would seem to be no risk of collision. Annex 10, Volume IV, Chapter 4 requires that these nuisance alerts be as infrequent as possible.
3.17.4.2 Compatible sense selection
The requirement in Annex 10, Volume IV, Chapter 4 is not intended to constrain the manner in which dangerous encounters are resolved, but rather is based on an appreciation that the majority of RAs are likely to be generated in encounters where there is no danger of collision. It places a statistical limit on the frequency with which ACAS disrupts ATC or the normal operation of the aircraft by inverting the vertical separation of two aircraft.
3.17.4.3 Deviations caused by ACAS
The restrictions on the deviations that may be caused by following RAs (as per Annex 10, Volume IV, Chapter 4) limit the disruption to normal aircraft operation as well as to ATC. While deviations from altitude clearances are the most obviously disruptive to ATC, other deviations, such as that caused by an RA to climb when the aircraft is descending, could be viewed equally seriously by ATC.
4.1.2.2 ACAS is designed to minimize the unnecessary alert rate when the standard vertical separation is 1 000 ft below FL 420. When a smaller separation standard is used, unnecessary RAs may occur; for example, when providing 500 ft separation against VFR traffic or allowing visual clearances through the level of other traffic.
4.2 INDEPENDENCE OF ACAS THRESHOLDS FROM ATC SEPARATION STANDARDS
ACAS thresholds are independent from ATC separation standards because ACAS does not strive to ensure separation, which is ATC’s role, but tries to avoid collision as a last resort.
5.5.4 Encounter Type 3 — Manoeuvring based on visual acquisition
5.5.4.1 A B747 and a DC10 were flying on converging tracks and both were mistakenly cleared to FL 370. When the controller detected the mistake, he attempted to reclear the DC10 to FL 350. In attempting to resolve this conflict, the controller used an incorrect callsign/flight number for the DC10.
5.5.4.2 The B747 pilot wrongly took the clearance meant for the DC10 and initiated a descent. At the same time, the B747 ACAS issued a Climb RA. However, the B747 pilot decided to not follow the RA because the B747 had visually acquired the DC10 and the descent was continued.
5.5.4.3 The DC10 pilot, who also had the B747 in sight, received a Descend RA that was followed. At the last moment, the DC10 pilot arrested the descent upon perceiving that the B747 was at the same altitude and also descending. Also at the last moment, the B747 pilot performed a sudden and violent escape manoeuvre that injured a number of passengers and flight attendants.
5.5.4.4 Because of the inappropriate manoeuvre based on visual acquisition, the B747 passed 10 metres below the DC10 with no lateral separation.
8.1.4 Calculation of risk ratio
8.1.4.1 Unresolved and induced risk of collision
8.1.4.1.1 When simulating the ACAS II logic on the generated encounters, two cases can occur which have an adverse contribution to the risk ratio:
a) the encounter presents a risk of collision and the ACAS II resolution fails to avoid it. This is an unresolved risk of collision; and
b) the encounter does not present a risk of collision and the ACAS II resolution creates it. This is an induced risk of collision.
8.1.4.1.2 Note that the induced risk of collision may become an important factor in the residual risk of collision. In other words, in airspaces in which ACAS has substantially reduced the overall risk of collision, when a collision does occur, it is possible that ACAS will be a causal factor in that collision.
10.3.5 General aviation monitoring
10.3.5.1 General aviation aircraft equipped with Mode A/C transponders were monitored for altitude reporting accuracy. Twenty-one of 548 transponders reported altitudes that deviated by more than 200 ft from the correct altitude. This reinforces the need for ongoing maintenance of transponders and monitoring of transponder performance while operating in the airspace.
Notwithstading the clear direction in Annex 11 [as quoted earlier], and considering the totality of Doc 9863, Aviation is far better off for having ACAS [particularly ACAS II and later variants] as a