In addition, during RNP AR approach procedures, the intermediate and final approach legs can include turns (Radius to Fix (RF) FMS legs) even shortly before DA (Decision Altitude). It has been envisaged that in some cases the DA will be reached during a turn. The RNP AR approach procedures are always designed to be flown with FMS vertical navigation based on Baro-VNAV principle.
Considering the more demanding environment of RNP AR, the authorities have developed more stringent regulations in addition to the MASPS ED-75/DO-236 of §3.2 and 3.3 above. These regulations involve certification but also operational requirements (FAA AC 90-101, OACI PBN Manual, AMC 20XZ).

3.5. ADDITIONAL NAVIGATION REQUIREMENTS FOR RNP AR 
As the obstacles can be located as close as a distance equal 2 times the RNP value, the probability to exceed this containment limit without annunciation must be lower than the  10-5/FH of MASPS ED-75/DO-236. All authorities have set the Target Level of Safety (TLS) at 10-7/procedure for this type of operations. The challenge is that the existing on board navigation systems (FMS with IRS, GPS updating and AP guidance) are not capable to achieve this target without operational mitigations. This is why a special authorization is required to ensure that operational procedures and pilot training will contribute at the adequate level to meet the expected target of level of safety. To achieve this safety objective at the aircraft design level alone would require a new design architecture similar to what we have for CATII or CATIII operations. But for the time being no aircraft manufacturer has designed such a system. The RNP AR operational concept has been developed to take the best advantage of existing system architecture complemented by the most efficient operational standards.  If an overall target level of safety of 10-7/procedure including the effect of failure cases cannot be demonstrated in certification alone without operational mitigation, the probability to exceed the containment limit at 2xRNP in normal conditions (without system or engine failure) can be demonstrated to be less than 10-7/procedure. This is computed with the statistical distribution of the TSE in the cross track direction for 
5.3 x. σ, (σ being the statistical standard deviation of the TSE distribution). We can see on the
drawing below that this condition is more  constraining than the accuracy requirement  at 1xRNP (95% of the time (~2x.
σ)).
..
.
TSE distribution in normal conditions

2σ
4.4σ 5.3σ 95% 1-10-5 1-10-7

σ
1xRNP

2xRNP
-11 -
3.  The RNP Concept
Getting to grips with RNP AR

To demonstrate this level of performance, in addition to the NSE, the FTE also needs to be
determined statistically based on flight and simulator tests.  The statistical determination of the
FTE has to consider the various conditions that may affect the flight path steering: tight turns,
　
中國航空網(wǎng) m.k6050.com
航空翻譯 www.aviation.cn
本文鏈接地址：getting to grips with RNP AR Required Navigation Performance(8)