RAIM prediction valid only for a preflight phase

[pilotnik]

Hey,

Why is RAIM prediction only valid for a preflight planning? If there is a NOTAM stating that RAIM is not available at the specific airport for a specific time, how come you are allowed to perform the RNP approach anyway?

Thank You

[OutsideCAS]

Why is RAIM prediction only valid for a preflight planning?


It might be to do with the on-board almanac update (happens roughly every 5 days?) not matching the more reliable and comprehensive data available 'real world' and available in a NOTAM, that's updated much more frequently and accurately, and thus required at the panning stage legally.


Perform the approach despite the RAIM prediction (NOTAM'd) saying N/A - I'd suggest this wouldn't be allowed for a purely GNSS approach. Depends on ground-based NAVAID's and all sorts of other considerations (WAAS?) also to allow approach other than purely GNSS...maybe?.

[RUMBEAR]

RAIM not available at planning means you need to plan contingencies such as ensure other approaches available. In flight the ability to commence an RNAV/RNP approach depends on actual nav system status. If RAIM actually not available you have alternative options.

[pilotnik]

Let me tell you why I am asking this. I am an ATCO and there was an RAIM unavailability prediction over the airport for a short time. Nevertheless, during this very time, TWO aircraft requested RNP approach. Both were informed about the lack of RAIM and even though they insisted on the approach.

[Check Airman]

The preflight RAIM check is just a prediction. It's possible that they did a "live" check at the start of the approach, and found that RAIM was available. The avionics will also let you know real time if RAIM is unavailable.

Not saying it was right- just raising a few possibilities.

What was the wx at the time? If it was VMC, they may have been able to fly the procedure without RAIM.

[Empty Cruise]

Pilotnik - you in the EU, your status says... now, if SBAS coverage is ok, you are fine without RAIM, so unless you also had an EGNOS outage, or your RNAV approach does not have LPV minima promulgated, the crew may be perfectly legal to fly the approach during a RAIM outage....

[latetonite]

A RAIM prediction is just a prediction. What you fly on is ANP. If you could get a better ANP than the RNP, you are good for the approach.

[FE Hoppy]

Pilotnik - you in the EU, your status says... now, if SBAS coverage is ok, you are fine without RAIM, so unless you also had an EGNOS outage, or your RNAV approach does not have LPV minima promulgated, the crew may be perfectly legal to fly the approach during a RAIM outage....

Exactly this.

We are LPV capable, use SBAS and have no requirement to check RAIM because the integrity is done through the SBAS signal.

[custardpsc]

pilotnik - what the others said - if the augmentation is satisfactory ( EGNOS in eu, WAAS in USA) then no need to check RAIM.

But I didn't get Empty Cruises point about LPV minima. I know that LPV requires WAAS but if you have integrity and there are no LPV minima what stops one using LNAV/VNAV approach / minima ?

[pilotnik]

The preflight RAIM check is just a prediction. It's possible that they did a "live" check at the start of the approach, and found that RAIM was available. The avionics will also let you know real time if RAIM is unavailable.

Not saying it was right- just raising a few possibilities.

What was the wx at the time? If it was VMC, they may have been able to fly the procedure without RAIM.

As far as I know there is no way to check RAIM availability on B737. The WX was IMC.

A RAIM prediction is just a prediction. What you fly on is ANP. If you could get a better ANP than the RNP, you are good for the approach.

OK, I was thinking about that too, but consider this: If RAIM is not available then the computer is not able to detect a bad signal from satellite. So if a bad signal cannot be detected the ANP will not reflect the actual performance at all. RNP approaches are ONLY allowed based on GNSS sensors, so it seems to me that without RAIM, the ANP reading is useless. Isn't that right?

pilotnik - what the others said - if the augmentation is satisfactory ( EGNOS in eu, WAAS in USA) then no need to check RAIM.

But I didn't get Empty Cruises point about LPV minima. I know that LPV requires WAAS but if you have integrity and there are no LPV minima what stops one using LNAV/VNAV approach / minima ?

I think you guys have a wrong idea on the RNP approaches, BUT it may also be that it is me who is wrong :rolleyes: From what I have read SBAS (or WAAS in USA) is only used for LPV approaches. It has nothing to do with LNAVs or LNAV/VNAVs. The only GNSS augmentation for the purpose of RNP APCH to the LNAV or LNAV/VNAV minimums is RAIM.

[Skyjob]

RAIM check can only be carried out by some operators preflight as aircraft are unable to detect it or access the (internet or ground) sources to retrieve it in flight.

If RAIM is therefore suitable preflight it can be used unless ATC advised to the contrary, they have access to last minute data.

In the OP follow up post he advises he is an ATCO and asked why aircraft asked to do such an approach when RAIM is not suitable.

Possibly the crew forgot to check preflight or perhaps their company manuals do not require them to do so (I doubt the latter).

If company does not include RAIM prediction for destination by default on the flight plan it usually requires additional crew actions preflight to access this data from another source, this can be time consuming and hinder an on time departure, or simply can be forgotten.

It would be recommended to advise operators to include this data for crew to read as standard during preflight, making them aware.

With RAIM outages far and few between with intervals not exceeding sometimes some minutes in EU, it is easy to miss an outage especially when they are printed in a graphical timeline of 3 days instead of numerical readout. Black and white printers producing the colour image also do not aid the crew in this. Just some random issues observed...

[Empty Cruise]

But I didn't get Empty Cruises point about LPV minima. I know that LPV requires WAAS but if you have integrity and there are no LPV minima what stops one using LNAV/VNAV approach / minima ?

Custard - you are of course right, if you are under SBAS coverage and intend to fly an RNP-1 STAR - no RAIM check required.

The point about promulgated LPV minima still stands (kinda) - because, if you are in an areas that should be covered by SBAS - why would you chose to not promulgate LPV minima? Obstacle clearance survey is basically the same as for LNAV/VNAV, so there is no money to save - which would make me think that something else is preventing a successful LPV implementation.

We operate in and out of SBAS coverage areas on a daily basis - so we have just decided to be done with it and always perform a PRAIM-check before departure, unless both crew agree that - yes, this is SBAS all the way.

In theory - SBAS ‘only’ saves one of the required 5 sats in visibility and decent geometry for FDE - if there is no GPS signal, SBAS is useless. So I do get your point - have been too lazy to consider it myself...

[underfire]

pilotnik.

RAIM prediction is used preflight, and if the RAIM prediction shows that RNP is not available enroute (depending on the procedure) or on final destination, legally, the flight cannot be planned to use the RNP procedure. (especially AR)

That is the simple fact of the flight plan, if RAIM prediction shows not available, you cannot plan to use the procedure.
We have been trying to change this, if conditions change enroute, but the criteria prevails.

Conversely, if RAIM prediction shows that it is avail, and during flight RAIM goes downhill on you, you cannot use the procedure...

We are LPV capable, use SBAS and have no requirement to check RAIM because the integrity is done through the SBAS signal

I hope you are kidding.

In general, reading through this thread on RAIM, the explanations and reasonings provided are sad, very, very sad.

[aterpster]

With TSO C145/146 (WAAS) avionics, if WAAS (SBS) is not available, FDE negates the need for a RAIM check, at least in FAA-dom.

[latetonite]

So, anybody starting an approach with: 'RAIM prediction from this morning checked, ANP checked'? And the actual RAIM, where you read it?
I would not mind starting an LNAV/VNAV approach, provided it was within my RNP, even if it was obtained with a sextant, so to speak.

[FE Hoppy]

undefire

Think you better take this up with BBD and the certifying authorities because FCOM 1 16-03-04 says the following:

"The SBAS performs integrity monitoring when the aircraft is operating in
SBAS coverage areas and the GNSS is operating in SBAS mode. When
the SBAS position is available, the predictive RAIM check is not required."

So no, I'm not kidding and the info I added could hardly be describes as sad. Accurate yes, sad not so much.

[A Squared]

OK, I was thinking about that too, but consider this: If RAIM is not available then the computer is not able to detect a bad signal from satellite. So if a bad signal cannot be detected the ANP will not reflect the actual performance at all. RNP approaches are ONLY allowed based on GNSS sensors, so it seems to me that without RAIM, the ANP reading is useless. Isn't that right?

No, it is not right. RAIM stands for Receiver *Autonomous* Integrity Monitoring. In plain English, that means the GPS receiver's ability to determine *by itself* the integrity of the GPS signals. The operative concept is *by itself* (that's what "autonomous" in the title means.) There are other means of verifying integrity and accuracy of GPS signals whcih are not "Receiver Autonomous". Systems of augmentation such as SBAS, WAAS, ground based augmentation, etc provide alternate means of determining the integrity of the GPS signals, so ensuring position integrity is not dependent on "RAIM". A NOTAM for lack of RAIM coverage does not make it illegal for all operators to fly an RNP approach. RNP only means "Required Navigation Performance" and there are different ways to achieve the required navigation performance for that approach. Without you having a detailed knowledge of the specific navigation equipment installed in the aircraft in question, and a knowledge of what was displayed on their cockpit navigation displays at the time, you have no way of knowing whether they were legal to fly the RNP approach. However, chances are pretty good that the were legal to fly the approach.

[oggers]

Let me tell you why I am asking this. I am an ATCO and there was an RAIM unavailability prediction over the airport for a short time. Nevertheless, during this very time, TWO aircraft requested RNP approach. Both were informed about the lack of RAIM and even though they insisted on the approach.

Very simply, the avionics will have indicated if they were able to carry out the procedure.

They may have been equipped with baro aiding which substitutes for one satellite or they may have been equipped with SBAS in which case RAIM is irrelevant. Even if they lacked either it is still quite feasible that the approach and the flight planning were totally legit The single piece of evidence “there was a predictive RAIM unavailability over the airport for a short time” is not in itself evidence of non compliance.

[pilotnik]

OK I think most of the answers are quite reasonable. A friend of mine just sent me some pics of A320 CDU and it indeed has live RAIM. I wasn't aware of that as on B737 there is no way to see anything more than ANP. I was trying to google what kind of GPS receiver do 737NGs have but I failed. Maybe there is an algorithm equivalent to RAIM inside but I can't find any info on it.

Another thing is that I really don't understand how could SBAS help doing RNP LNAV or LNAV/VNAV approach? Or why would baro aiding allow for one satellite less in case of LNAV/VNAV? For LNAV, baro aid is obsolete. For LNAV/VNAV baro aid is a must and if it allowed for a satellite number reduction it would mean that reduced number of satellites is in fact a standard number... And for LPV? This is the only, fully RNP GNSS approach and baro aiding here is useless.

I think some of you guys confuse ABAS, SBAS and even GBAS. I've read through dozens of documents on PBN and interchangeable use of augmentation types is nowhere to be found. You can't just use GBAS or SBAS to compensate for lack of satellites in order to perform LNAV/VNAV. Am I crazy or am I missing something important here?

[A Squared]

Well, as far as the RAIM and baro-aiding, Here's how RAIM works on a conceptual, mathematical level. A GPS receiver computes a position by solving for 4 unknowns: Coordinates in 3 Dimensions (X,Y,Z) and receiver time offset from the GPS satellite system. The latitude and Longitude you see displayed on the receiver are translated from the GPS systems native geocentric (earth centered) coordinate system. As you may remember from algebra, to solve a system of equations for 4 unknowns, you have to have 4 equations. Each measurement to a different satellite adds one equation to the system, so if you have measurements to 4 satellites, the receiver can solve for X,Y,Z and dT. However, with only 4 satellites you have no way of checking that the 4 signals are all correct. However if you have measurements to 5 satellites you have a redundant, or over-determined solution, and with a redundant solution you can tell if the measurement from one satellite doesn't fit the solution. But you have no way of telling which measurement is bad, just that the 5 measurements don't all fit the solution within a certain tolerance. If you have measurements to 6 satellites, that gives you a tie breaker, which allows you to determine which satellite signal is bad, and drop it from the solution.

Now, how does baro-aiding fit into that? Baro-aiding can be used to add one more observation to the solution. If you have baro-aiding, you know what your altitude is, and that essentially reduces the number of unknowns that you have.

If it helps, you can think of it as the receiver solving for lat, long, altitude, and time offset, requiring 4 satellites for a solution, but with baroaiding, altitude is known, so you can solve for lat, long, and time offset with measurements from 3 satellites. That's not quite accurate, like I said the receiver is calculating in Geocentric Cartesian coordinates, but conceptually it's close enough that it's not incorrect, exactly. This was common in the early days of GPS, especially for marine applications, when the GPS satellite system was not yet complete and coverage was spotty. GPS receivers had an option of the user entering the elevation if known (sea level plus antenna height for a boat on the ocean ) which allowed navigation on only 3 GPS satellites. Baro-aiding does the same thing, it automatically enters the altitude derived from the barometric altimeter to the GPS solution and reduces the number of satellites required by 1; 3 satellites fro a minimum position solution, 4 satellites for an over determined position that checks for a good solution, and 5 satellites for the minimum satellites to catch an error, and identify the satellite signal in error.

So, that's the basic concept. I would expect that the TSOs and other standards under whcih the systems are designed and approved would require a higher level of redundancy, so it's likely that "RAIM Availability" from a regulatory standpoint requires more satellites than in my description, but at the algebra/geometry level, that's what's required.

[A Squared]

I think some of you guys confuse ABAS, SBAS and even GBAS. I've read through dozens of documents on PBN and interchangeable use of augmentation types is nowhere to be found. You can't just use GBAS or SBAS to compensate for lack of satellites in order to perform LNAV/VNAV. Am I crazy or am I missing something important here?

You're still stuck on "Do we have RAIM" The requirement for a RNP approach (or any RNP operation) is not; "do we have RAIM", it is; "Do we have the required navigation performance" My FMS system (SBAS/WAAS capable GPS receivers and triple IRU's ) doesn't tell me RAIM/No RAIM, (although I can find that if I'm interested) it tells me what the current navigation performance is, and if that is more accurate the the minimum accuracy required for the operation, then we're good to go. The system as a whole is not dependent on RAIM to determine the current navigation performance. there are other ways that it automatically verifies the performance, and augmentation is one of the ways it does that.

[pilotnik]

OK, that is logical and I think it is more clear to me now. This thread and you guys helped me a lot. Thank you for your time. :ok:

[NICOOOLDUDU]

A Squared, I think you are correct.

Now one question. Let say you have an integrated navigation system (like on the airbus) that computes your position and accuracy (and also your Required Navigation Performance), by using your gps in priority but also Dme, vor, ndb and dead reck with the intertials if you are in big ****.

Now you loose your 2 gps. However you still receive dme and/or vor coverage and your navigation performance is still computed and (theoretically) within limits for approach. Would it mean you can still start your RNAV approach?

The next day, you loose not only your 2 gps but also all your antenas that include vor/dme/adf just approach, bad day...luckily your fms is still serviceable to calculate your actual and required navigation performance for an RNAV approach...now your 2-3 irs become your primary means of navigation, but because of short time they didn't drifted much and your nav perf shows "good to go"...so do you do your RNAV approach?

Thanks for your feedback, that's an interesting topic.

[FE Hoppy]

If you read the specification for RNP APCH your question is answered.

[EMIT]

A Squared beat me with the provision of a pretty good explanation, but I'll post this anyway.



The question of the O.P. is a good one and in trying to answer it correctly we do see the limits of technical knowledge that is supplied to us, operators, via FCOM’s, Ops Manuals, etcetera.

We can only hope that the guys writing the books were correct in every detail.

Let’s look at GNSS basics – a receiver needs signals from 4 satellites in order to be able to compute a position. If the signal from a satellite is flawed, the computed position will be wrong.
If the receiver can track more than the minimum number of 4 satellites, it can use the multitude of acquired ranges to check whether all the different combinations of ranges yield the same position as a result. If all the position computations with the different sets of ranges give (nearly) identical position results, then the position has integrity, is reliable.
If the receiver can track 5 satellites, and one satellite signal is flawed, then the integrity check will fail.
If the receiver can track 6 satellites, it can detect a flawed position and it can determine which satellite signal is flawed and can exclude that signal from the process (FDE, Fault Detection and Exclusion).
If the receiver can track even more satellites, more redundancy is available.

If a GNSS receiver cannot supply a position, it is required to provide an alert to the pilot – what the pilot books do not properly explain, is exactly what conditions would trigger that alert.
Of course, if the GNSS set would be completely dead, due to receiving no electrical power, it is obvious that it would be declared faulty; however, whether it supplies a failure flag when, for instance, it cannot see enough satellites in order to perform the integrity check, that is not explained.
It would be logical if the GNSS position would be declared unusable if it was made up of only 4 satellite signals, but we, as pilots, just do not know whether the sets work that way.

In order to receive a usable signal, a receiver (antenna) needs line of sight to a satellite.
Since the satellites orbit relatively low above the earth, they are not geostationary, therefor the visibility of satellites to receivers changes continuously.
It is possible, from knowledge of the satellite orbits, coupled with announced outages of satellites, to compute whether at a certain position, at a certain time, there will be enough operational satellites visible to a receiver, for performance of RAIM. Such a computation, named RAIM prediction, can be made for an entire route of flight.
If a receiver only has RAIM, visibility of 5 satellites would be required. If the receiver is fancier (capable of FDE), 6 satellites would be required, and the prediction algorithm would be set for that requirement.

Regulations require that, for portions of a flight, where GNSS is required for RNP, during flight planning, a check is made of availability of enough satellites for RAIM. If gaps of more than 5 minutes are predicted, replanning is required, either via route adjustment, or by time adjustment of the flight.

Once in flight, there is no longer a requirement to pay attention to the RAIM availability predictions, only in case of inflight replanning, there could be a requirement that the operator’s dispatch unit recalculates a prediction (from a regulatory viewpoint, inflight re-planning is more or less equivalent to preflight planning). For the actual flight execution, the only thing important is that the performance of the NAV system of the aircraft keeps Actual Navigation Performance (ANP) within the bounds of Required Navigation Performance (RNP).

Our OM-A literally states: An in-flight RAIM availability check is not required prior to commencing the procedure. This should answer question of OP for my specific company and aircraft type.
As a pilot we can only assume, for lack of technical knowledge, that then the GPS receiver must annunciate a failure if it does not receive enough satellites to perform RAIM.

As a further note about the predictions: we are required to use the FDE algorithm for the prediction, and the mask angle is set at 5 degrees. So, the prediction tells us whether at least 6 operational satellites will be higher than 5 degrees above the horizon. As explained above however, already 5 satellites are enough to determine whether the derived position is reliable or not. Also, a satellite at 4.95 degrees above the horizon might provide a usable signal, but it would not have counted in the prediction. So the prediction is conservative and in real life GNSS might work just fine even though the prediction would have said NO.

All specific remarks about equipment and regulations will depend on equipment actually built into aircraft and ops manuals of company.

As a final note, just to be sure, RAIM is not a sort of special signal that has to be sent by satellites or has to be received in any way, it is a check, internally in the GNSS receiver, that can be made when more operational satellites can be received simultaneously then basically required for computation of position.
The receiver will track as many satellites as possible (up to the maximum capability of the receiver) and it should be simple to provide an alert if the number of tracked satellites is less than required for performance of RAIM to confirm the integrity of position.

The computation of ANP is another area where our documentation is very sparse with details – the NAV set computes and displays ANP and alerts the crew when ANP exceeds RNP, that’s as deep as it goes.

There are a couple of websites that provide technical principles, but detailed info about specific equipment in one’s aircraft is hard to get.
ESA (European Space Agency) Navipedia.net
US GPS site gps.gov

[LeadSled]

With TSO C145/146 (WAAS) avionics, if WAAS (SBS) is not available, FDE negates the need for a RAIM check, at least in FAA-dom. Folks,
Does this post from aterpster not really define the situation.

You must know the certified capabilities of what you have on board, and its operational limitations.

I can well imagine that some authorities have a pre-flight planning requirement related to RAIM predictions, but RAIM is not applicable to later generation GPS receivers. If you have C-129 based equipment RAIM is applicable in flight. This is so, whether or not you have WAAS/SBAS for a particular area, WITH c-145/146 BASED EQUIPMENT, all WAAS/SBAS availability does is change the minima.