4.2 Evaluation of the extrapolation at the display time
One method could be the computation of the accuracy of the displayed tracks. But this implies that the reconstruction of the reference trajectories at the display time is possible. This is not the case with CENA tools for the time being. This could be possible with the use of MURATREC but, as already mentioned, France has no experience with (or no confidence in) this RASS software yet.
Another method could be the computation of the accuracy of the speed vector elaborated by the tracker. This implies that the reconstruction of the reference trajectories is good both in position and speed vector. As far as CENA home tools are concerned, this is not the case. The accuracy of the reconstructed speed vector is about the same order of magnitude as the one of the tracker itself, as soon as Multi Plot Variable Update (MPVU) techniques are used for multi-radar tracking.
So, in the absence of the appropriate tool, an approximate method is used for the time being : Extrapolations at the display time are compared with the positions of the reference plots corresponding to the updates out of the tracker used for the extrapolations. In this method, the positions which are compared do not correspond in time. So this method does not allow the calculation of standard deviation for the errors of the displayed positions. Nevertheless, the distribution of the deviations is studied and gives some representative clues for the quality of the extrapolation process.
Special attention is put on :
-the cases already found in the evaluation of the tracker (residual errors greater than x NM) so as to check how these situations are dealt with by the extrapolation process.
-all the cases on the tail of the distribution (errors greater than y NM 1)
The percentage of residual errors must not be greater than that of the mono-radar case.
4.3 Evaluation of the maximum of the errors in the co-ordinate transform of targets with no mode C
As mentioned in chapter 3, the errors induced in the co-ordinate transform by the slant range correction in case of total absence of mode C can not be neglected. Indeed, these errors can not be completely avoided as soon as the system is not informed at all of the aircraft flight level. So they have to be minimised so as not to waste the accuracy wealth of the tracker.
In France this problem is mainly studied with simulations, by comparison, for a given scenario, of the positions shown on the display when the flight level is known and the same positions obtained with a false estimation of the flight level.
1)
We use the same parameter as the one used for the evaluation of the tracker (x NM).
A scenario consists in assuming a given error in the estimation of the level and make the position of the target vary in relation with the position of the centre of projection for the co-ordinate transform.
As the results are site dependant, this is done for every implementation of a MRDPS in a new radar environment. If needed, the flight level estimation process used within the MRDPS is tuned and a test is done with real traffic so as to check that the maximum of the errors is never greater than 1 NM.
ANNEX 2 Attachment C
EXAMPLE METHODOLOGY FOR MULTI-RADAR TRACKING SYSTEMS
Minimum Radar Separation of 5 NM, used at the Maastricht UAC (Methodology also used by Austro Control)
1. GENERAL CONSIDERATIONS
中国航空网 www.aero.cn
航空翻译 www.aviation.cn
本文链接地址：GUIDELINES FOR THE APPLICATION OF THE ECAC RADAR SEPARATION(25)
