Help getting highest INS accuracy for pre-planned popups

[Hatman335]

Before we're getting seriously sidetracked with JDAMs and car GPS and whatnot, we should establish a baseline. What is the intended CEP for a GPS solution in the F-16 module? Is the ~200 feet offset that OP is seeing intentional? @Lord Vader Can you please clarify? 

[Lord Vader]

Hi, according to the publicly available documentation the CEP for the GBU-31/38 JDAM is 5 meters. We have modelled this accordingly in DCS. 

Just be aware that currently the JDAM self-guidance/correction is WIP so if you drop a bomb without a sensor track (FCR, TGP), the bomb will only get the information the MMC sends, which included INS drift. 

[Hatman335]

Just now, Lord Vader said:

Hi, according to the publicly available documentation the CEP for the GBU-31/38 JDAM is 5 meters. We have modelled this accordingly in DCS. 

 

Thank you for the answer, but let me please clarify, since this wasn't what I was trying to ask. I'm asking about the CEP (or error value, or inaccuracy) in the navigation solution when using GPS.

 

Assuming you place a steerpoint at a given location, you have a good GPS solution, you don't do any fixes or updates, you don't introduce any system deltas and just fly, there how close is it intended to remain to the selected location? What level of offset or inaccuracy should you observe when using GPS and what factors influence this in DCS?

 

Thank you in advance for the response.

[Lord Vader]

Oh I see,

If you consult our white paper, you can see that the Kalman filter output error can get to about 15m.

https://www.digitalcombatsimulator.com/upload/medialibrary/941/fsjcbyvt707ib9q7s7szpsra373gxak7/F-16_INS_White_Paper.pdf

[Hatman335]

22 minutes ago, Lord Vader said:

If you consult our white paper, you can see that the Kalman filter output error can get to about 15m.

 

I could be reading this wrong, but it seems to me that picture 4 on page 6 shows that a blended INS+GPS solution in a reasonable time period (say within an hour of flight time) could still have up to 60 meters of error. If that is the case, then OP's findings with approximately 200 feet of drift may be intended?

[Lord Vader]

That was a demonstration over a long period of time. This could be caused by various factors that were not explained in that graphic. Basically, manoeuvring and acceleration can influence the deviations, but you'll notice the filter will eventually reduce the deviation over time.

[Hatman335]

1 minute ago, Lord Vader said:

That was a demonstration over a long period of time

The maximum error in that graph is reached at 2500 seconds, which less than an hour. So in the current implementation is it recommended to have enough straght and level flying time prior to a system delivery so that the solution can be corrected by the filter? 

[Lord Vader]

Again, it's not explained what happened in that graphic. I am unsure if, for example, there were some extreme accelerations or GPS suffered some sort of interference for the purpose of demonstrating degradation. Under normal circumstances, yes, if there is a degradation over time, the Kalman filter will continuously try to remove the maximum amount of errors possible. 

[Hatman335]

Okay, so if I understand everything correctly, that level of drift would be highly unusual and only due to some kind of extreme factor. I assume that normal dynamic flying would not be sufficient to cause that level of error, therefore if all works well, OP should see a better accuracy/lower error value than this.

[Hatman335]

I've also done two tests and the offset can definitely reach up the 200~ feet range. 

[PawlaczGMD]

12 hours ago, Lord Vader said:

Again, it's not explained what happened in that graphic. I am unsure if, for example, there were some extreme accelerations or GPS suffered some sort of interference for the purpose of demonstrating degradation. Under normal circumstances, yes, if there is a degradation over time, the Kalman filter will continuously try to remove the maximum amount of errors possible. 

And to be absolutely clear, does taking a FIX (e.g. with the TGP) restore the accuracy?

E.g. If I'm flying to drop a JDAM on GPS coordinates without a visual, and it will take me 40 mins to get there, if I take a FIX 5 minutes before the drop, will I hit with CEP ~5m? 

10 hours ago, Hatman335 said:

I've also done two tests and the offset can definitely reach up the 200~ feet range. 

I've also experienced similar degradation after 30-40 mins of gentle flight.

[Nealius]

IIRC JDAM "drift" is a separate issue where JDAM takes into account the jet's INS drift instead of using guidance from the GPS satellites once released.

This thread is about INS accuracy itself and if the fixes work for old-school weapons delivery. "Fixing" the JDAM accuracy could be a byproduct of this but I'm not sure that it's the cause.

Edited October 23, 2024 by Nealius

[PawlaczGMD]

1 hour ago, Nealius said:

IIRC JDAM "drift" is a separate issue where JDAM takes into account the jet's INS drift instead of using guidance from the GPS satellites once released.

This thread is about INS accuracy itself and if the fixes work for old-school weapons delivery. "Fixing" the JDAM accuracy could be a byproduct of this but I'm not sure that it's the cause.

 

I would also naively think that a JDAM dropped on a STPT doesn't need to interact with the jet's INS at all, as it has its own GPS guidance that should be good to 5 meters, but it's not the case. The drop solution could be affected, but a JDAM has more than enough room  to fix even a 200 foot offset when dropped from high altitude.

Either way, I am also just interested in how to fix this in practice right now.

[itn]

  

14 minutes ago, PawlaczGMD said:

I would also naively think that a JDAM dropped on a STPT doesn't need to interact with the jet's INS at all, as it has its own GPS guidance that should be good to 5 meters, but it's not the case. The drop solution could be affected, but a JDAM has more than enough room  to fix even a 200 foot offset when dropped from high altitude.

Either way, I am also just interested in how to fix this in practice right now.

So one more time: For now you need to give the JDAMs a sensor track (i.e. SPI from TGP) to have any sort of accuracy with them. Nothing much else you can do, until ED brings back the feature(s) that makes JDAM a JDAM. Forget what you know or you think you know about JDAMs, the current DCS implementation is not that.

This was already answered in this thread, in your thread on Sunday, and at least one if not more previous threads about the JDAM issues.

 

 

 

 

 

[SpecterDC13]

Also, dont forget that when using the TGP you need to TMS UP long. That will send the SPI data to the JDAM. 

[Tenkom]

19 hours ago, SpecterDC13 said:

Also, dont forget that when using the TGP you need to TMS UP long. That will send the SPI data to the JDAM. 

Are you sure you're not thinking about the A10c? I didn't know tms-up long is a thing in the viper.

[Default774]

19 hours ago, SpecterDC13 said:

Also, dont forget that when using the TGP you need to TMS UP long. That will send the SPI data to the JDAM. 

I think you are confusing the viper with another aircraft. You have to do no such thing in the viper

[falconzx]

SPI is constantly updated (and sent to bomb) by the Sensor Of Interest.

[Hatman335]

This issue is not limited to JDAMs, we need to figure out why this apparent discrepancy between the intended and the measured accuracy exists. In fact this mostly affects system deliveries with unguided weapons.

[gonvise]

It's probably my fault, but in missions where the GBU-38 is used and the objectives have coordinates, such as missions 6 or 7, once these are entered into a WP during the trip to the area, when arriving and checking the objective with the TGP the coordinates are not above the objective, but a few meters off. Is there currently a bug regarding this or am I doing something wrong?

[baltic_dragon]

I think this is a bug. I've extracted all the coordinates and elevations directly from the mission editor... So no idea why this happens. 

[Pillowcat]

Known behavior: plane nav system drift affects carried gps bombs. No one knows would it be changed anytime soon or ever.

 

[BIGNEWY]

threads merged

[gonvise]

@BIGNEWY, can you confirm if this bug is being looked into or if there is some workaround for it? Thanks.

[Lord Vader]

First of all, there are two separate items in this thread, one related to the other, but not the same. One has to do with how the INS+GPS system works using the Kalman filter. Another is the information IDMs receive from the MMC and the ability of the ammunition to self-correct, which is not really related to the F-16C INS.

Neither are "bugs", one is a possible deviation on the Kalman filter logics and we're testing the data and will update the logics if needed. The other is a feature not yet available in IDM weapons we're working to implement in the near future.