Landing predictions become inf / nan

[KevWal]

Playing back this GPS track: https://github.com/KevWal/pits/blob/63f0a1918b78d8278c807f693a08f81b01c122c2/nmea/Friday-No-GSV-VTG.NMEA

I get issues with the prediction code:

LORA0: $$KW01,236,21:03:57,52.01939,-0.15447,17022,35,83,5,31.3,4.317,0.99,52.29643,2.62463,3.3,3333,-70,6,22,NBA6C
...
LORA0: $$KW01,237,21:04:16,52.01968,-0.15169,16838,1,0,5,31.2,4.273,0.97,52.29172,2.58097,3.3,3270,-64,8,23,N8479
...
LORA0: $$KW01,238,21:04:27,52.01996,-0.14920,16674,35,83,5,31.2,4.356,inf, nan, nan,0.0,2147483647,-64,8,23,N1051
...
LORA0: $$KW01,239,21:04:27,52.02020,-0.14700,16529,35,77,5,31.2,4.311,inf, nan, nan,0.0,2147483647,-64,8,23,NBCD6

Playing it back a second time, I get the same thing (Scroll back buffer didnt go back far enough to find the start of the issue):

LORA0: $$KW01,254,22:39:14,52.03025,0.02985,11048,87,91,5,30.4,4.350,inf, nan, nan,0.0,2147483647,-62,6,64,N6B11
...
LORA0: $$KW01,255,22:39:30,52.03013,0.03556,10942,87,91,5,30.6,4.276,inf, nan, nan,0.0,2147483647,-62,6,64,N0E80
...
LORA0: $$KW01,256,22:39:47,52.02990,0.04163,10833,87,92,5,30.6,4.353,inf, nan, nan,0.0,2147483647,-62,6,64,N*3319

Interestingly, 2 values are fixed again at the very end of the playback:

LORA0: SSDV image 176 packet 79 of 114
LORA0: $$KW01,408,23:19:34,51.99414,0.58410,00126,20,92,5,31.0,4.346,inf, nan, nan,0.0,2147483647,-60,6,95,NC940
LORA0: SSDV image 176 packet 80 of 114
LORA0: $$KW01,409,23:19:35,51.99412,0.58437,00116,20,92,5,31.0,4.263,inf, nan, nan,0.0,2147483647,-60,6,95,N2BA4
LORA0: SSDV image 176 packet 81 of 114
LORA0: $$KW01,410,23:19:35,51.99412,0.58454,00109,20,92,5,31.0,4.263,inf, nan, nan,0.0,2147483647,-60,6,95,ND6F9
LORA0: ^^KW01,434.712,1,2,48,96,007DE
LORA0: SSDV image 176 packet 82 of 114
LORA0: $$KW01,411,23:19:45,51.99410,0.58506,00090,14,90,5,30.9,4.263,inf,51.99411,0.58481,0.0,0,-60,6,95,NAA24
LORA0: SSDV image 176 packet 83 of 114
LORA0: $$KW01,412,23:19:47,51.99410,0.58519,00084,1,0,5,30.9,4.300,inf,51.99410,0.58519,0.0,0,-60,6,95,N1E58
LORA0: SSDV image 176 packet 84 of 114
LORA0: $$KW01,413,23:19:53,51.99410,0.58560,00065,14,90,5,30.9,4.300,inf,51.99410,0.58519,0.0,0,-60,6,95,N*FFFA

Cheers
Kev

[PiInTheSky]

When you say "Playing it back a second time", are you replacing the GPS file while the tracker program is running, or is this a separate run of the tracker?

…

On Tue, 23 Aug 2022 at 07:15, Kevin Walton ***@***.***> wrote: Playing back this GPS track: https://github.com/KevWal/pits/blob/63f0a1918b78d8278c807f693a08f81b01c122c2/nmea/Friday-No-GSV-VTG.NMEA I get issues with the prediction code: LORA0: $$KW01,236,21:03:57,52.01939,-0.15447,17022,35,83,5,31.3,4.317,0.99,52.29643,2.62463,3.3,3333,-70,6,22,N *BA6C ... LORA0: $$KW01,237,21:04:16,52.01968,-0.15169,16838,1,0,5,31.2,4.273,0.97,52.29172,2.58097,3.3,3270,-64,8,23,N* 8479 ... LORA0: $$KW01,238,21:04:27,52.01996,-0.14920,16674,35,83,5,31.2,4.356,inf, nan, nan,0.0,2147483647,-64,8,23,N *1051 ... LORA0: $$KW01,239,21:04:27,52.02020,-0.14700,16529,35,77,5,31.2,4.311,inf, nan, nan,0.0,2147483647,-64,8,23,N*BCD6 Playing it back a second time, I get the same thing (Scroll back buffer didnt go back far enough to find the start of the issue): LORA0: $$KW01,254,22:39:14,52.03025,0.02985,11048,87,91,5,30.4,4.350,inf, nan, nan,0.0,2147483647,-62,6,64,N *6B11 ... LORA0: $$KW01,255,22:39:30,52.03013,0.03556,10942,87,91,5,30.6,4.276,inf, nan, nan,0.0,2147483647,-62,6,64,N*0E80 ... LORA0: $$KW01,256,22:39:47,52.02990,0.04163,10833,87,92,5,30.6,4.353,inf, nan, nan,0.0,2147483647,-62,6,64,N*3319 Interestingly, 2 values are fixed again at the very end of the playback: LORA0: SSDV image 176 packet 79 of 114 LORA0: $$KW01,408,23:19:34,51.99414,0.58410,00126,20,92,5,31.0,4.346,inf, nan, nan,0.0,2147483647,-60,6,95,N *C940 LORA0: SSDV image 176 packet 80 of 114 LORA0: $$KW01,409,23:19:35,51.99412,0.58437,00116,20,92,5,31.0,4.263,inf, nan, nan,0.0,2147483647,-60,6,95,N*2BA4 LORA0: SSDV image 176 packet 81 of 114 LORA0: $$KW01,410,23:19:35,51.99412,0.58454,00109,20,92,5,31.0,4.263,inf, nan, nan,0.0,2147483647,-60,6,95,N *D6F9 LORA0: ^^KW01,434.712,1,2,48,96,0*07DE LORA0: SSDV image 176 packet 82 of 114 LORA0: $$KW01,411,23:19:45,51.99410,0.58506,00090,14,90,5,30.9,4.263,inf,51.99411,0.58481,0.0,0,-60,6,95,N *AA24 LORA0: SSDV image 176 packet 83 of 114 LORA0: $$KW01,412,23:19:47,51.99410,0.58519,00084,1,0,5,30.9,4.300,inf,51.99410,0.58519,0.0,0,-60,6,95,N* 1E58 LORA0: SSDV image 176 packet 84 of 114 LORA0: $$KW01,413,23:19:53,51.99410,0.58560,00065,14,90,5,30.9,4.300,inf,51.99410,0.58519,0.0,0,-60,6,95,N*FFFA Cheers Kev — Reply to this email directly, view it on GitHub <#49>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4ERK3OYHIVYKBQOINNI3DV2RT67ANCNFSM57KD4LXQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[PiInTheSky]

Can you send me your NMEA file please, and I'll set this up today. Dave

…

On Tue, 23 Aug 2022 at 07:15, Kevin Walton ***@***.***> wrote: Playing back this GPS track: https://github.com/KevWal/pits/blob/63f0a1918b78d8278c807f693a08f81b01c122c2/nmea/Friday-No-GSV-VTG.NMEA I get issues with the prediction code: LORA0: $$KW01,236,21:03:57,52.01939,-0.15447,17022,35,83,5,31.3,4.317,0.99,52.29643,2.62463,3.3,3333,-70,6,22,N *BA6C ... LORA0: $$KW01,237,21:04:16,52.01968,-0.15169,16838,1,0,5,31.2,4.273,0.97,52.29172,2.58097,3.3,3270,-64,8,23,N* 8479 ... LORA0: $$KW01,238,21:04:27,52.01996,-0.14920,16674,35,83,5,31.2,4.356,inf, nan, nan,0.0,2147483647,-64,8,23,N *1051 ... LORA0: $$KW01,239,21:04:27,52.02020,-0.14700,16529,35,77,5,31.2,4.311,inf, nan, nan,0.0,2147483647,-64,8,23,N*BCD6 Playing it back a second time, I get the same thing (Scroll back buffer didnt go back far enough to find the start of the issue): LORA0: $$KW01,254,22:39:14,52.03025,0.02985,11048,87,91,5,30.4,4.350,inf, nan, nan,0.0,2147483647,-62,6,64,N *6B11 ... LORA0: $$KW01,255,22:39:30,52.03013,0.03556,10942,87,91,5,30.6,4.276,inf, nan, nan,0.0,2147483647,-62,6,64,N*0E80 ... LORA0: $$KW01,256,22:39:47,52.02990,0.04163,10833,87,92,5,30.6,4.353,inf, nan, nan,0.0,2147483647,-62,6,64,N*3319 Interestingly, 2 values are fixed again at the very end of the playback: LORA0: SSDV image 176 packet 79 of 114 LORA0: $$KW01,408,23:19:34,51.99414,0.58410,00126,20,92,5,31.0,4.346,inf, nan, nan,0.0,2147483647,-60,6,95,N *C940 LORA0: SSDV image 176 packet 80 of 114 LORA0: $$KW01,409,23:19:35,51.99412,0.58437,00116,20,92,5,31.0,4.263,inf, nan, nan,0.0,2147483647,-60,6,95,N*2BA4 LORA0: SSDV image 176 packet 81 of 114 LORA0: $$KW01,410,23:19:35,51.99412,0.58454,00109,20,92,5,31.0,4.263,inf, nan, nan,0.0,2147483647,-60,6,95,N *D6F9 LORA0: ^^KW01,434.712,1,2,48,96,0*07DE LORA0: SSDV image 176 packet 82 of 114 LORA0: $$KW01,411,23:19:45,51.99410,0.58506,00090,14,90,5,30.9,4.263,inf,51.99411,0.58481,0.0,0,-60,6,95,N *AA24 LORA0: SSDV image 176 packet 83 of 114 LORA0: $$KW01,412,23:19:47,51.99410,0.58519,00084,1,0,5,30.9,4.300,inf,51.99410,0.58519,0.0,0,-60,6,95,N* 1E58 LORA0: SSDV image 176 packet 84 of 114 LORA0: $$KW01,413,23:19:53,51.99410,0.58560,00065,14,90,5,30.9,4.300,inf,51.99410,0.58519,0.0,0,-60,6,95,N*FFFA Cheers Kev — Reply to this email directly, view it on GitHub <#49>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4ERK3OYHIVYKBQOINNI3DV2RT67ANCNFSM57KD4LXQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[KevWal]

Hi, the 'second run' was a totally independent run of the tracker code. The tracker code exits at the end of the NMEA file, and I re run the program from the command line. Fyi, I had the same issue on a different set of test data last night, I will re run that today and capture it in a new comment here. Thanks Kev

…

On 24 August 2022 08:19:33 BST, Pi In The Sky Project ***@***.***> wrote: When you say "Playing it back a second time", are you replacing the GPS file while the tracker program is running, or is this a separate run of the tracker?

[KevWal]

Can you send me your NMEA file please, and I'll set this up today. Dave

Hi Dave,

File attached (I had to add .txt at the end to let me upload it), or also linked on gihub in the mesage above.

Cheers
Kev

Friday-No-GSV-VTG.NMEA.txt

[KevWal]

Second example (https://github.com/KevWal/pits/blob/develop/nmea/XABEN11.NMEA) re running, currently it is on the way up, it broke on the way down here last night:

(where the descent becomes vertical)

[KevWal]

Just caught the break before it scrolled out of the buffer! :)

LORA0: $$KW01,690,12:43:13,52.06396,0.65472,24148,3,0,5,28.1,4.356,1.29,51.96360,1.03043,2.9,4435,0,0,0,N3E5E
LORA0: $$KW01,691,12:43:38,52.06362,0.65289,23761,0,0,5,28.0,4.305,1.75,51.94694,1.09340,2.5,5149,0,0,0,NACC0
LORA0: $$KW01,692,12:43:54,52.06353,0.65208,23532,0,0,5,28.1,4.199,1.67,51.94973,1.08321,2.5,5016,0,0,0,N6A25
LORA0: $$KW01,693,12:44:18,52.06343,0.65094,23189,22,265,5,28.3,4.366,inf, nan, nan,0.0,2147483647,0,0,0,NB52C
LORA0: $$KW01,695,12:44:51,52.06330,0.64941,22732,22,265,5,28.4,4.252,inf, nan, nan,0.0,2147483647,0,0,0,N8601
LORA0: $$KW01,696,12:45:11,52.06321,0.64857,22427,0,0,5,28.3,4.363,inf, nan, nan,0.0,2147483647,0,0,0,N4276

Will leave it running as yesterday it fixed its self at the very end.

Cheers
Kev

[PiInTheSky]

The inf is CDA which implies zero descent rate, so the issue will be somewhere around the use of altitude. I ran my NMEA twice, no issues. Currently running XABEN11. Dave

…

On Wed, 24 Aug 2022 at 14:03, Kevin Walton ***@***.***> wrote: Just caught the break before it scrolled out of the buffer! :) LORA0: $$KW01,690,12:43:13,52.06396,0.65472,24148,3,0,5,28.1,4.356,1.29,51.96360,1.03043,2.9,4435,0,0,0,N *3E5E LORA0: $$KW01,691,12:43:38,52.06362,0.65289,23761,0,0,5,28.0,4.305,1.75,51.94694,1.09340,2.5,5149,0,0,0,N* ACC0 LORA0: $$KW01,692,12:43:54,52.06353,0.65208,23532,0,0,5,28.1,4.199,1.67,51.94973,1.08321,2.5,5016,0,0,0,N *6A25 LORA0: $$KW01,693,12:44:18,52.06343,0.65094,23189,22,265,5,28.3,4.366,inf, nan, nan,0.0,2147483647,0,0,0,N*B52C LORA0: $$KW01,695,12:44:51,52.06330,0.64941,22732,22,265,5,28.4,4.252,inf, nan, nan,0.0,2147483647,0,0,0,N *8601 LORA0: $$KW01,696,12:45:11,52.06321,0.64857,22427,0,0,5,28.3,4.363,inf, nan, nan,0.0,2147483647,0,0,0,N*4276 Will leave it running as yesterday it fixed its self at the very end. Cheers Kev — Reply to this email directly, view it on GitHub <#49 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4ERK7DGUINXWGVBAD2RQ3V2YMQRANCNFSM57KD4LXQ> . You are receiving this because you commented.Message ID: ***@***.***>

[KevWal]

Want to send me your NMEA and I can run here once https://github.com/KevWal/pits/blob/develop/nmea/XABEN11.NMEA is finished?

[KevWal]

XABEN11 now finished here, and as per my run last night, the nan's worked at the very end, but not the inf:

LORA0: $$KW01,927,13:47:06,51.98930,0.95826,00109,0,0,5,31.0,4.246,inf, nan, nan,0.0,2147483647,0,0,0,NE15C
LORA0: $$KW01,928,13:47:09,51.98926,0.95841,00103,39,103,5,31.0,4.246,inf, nan, nan,0.0,2147483647,0,0,0,NAE8B
LORA0: $$KW01,929,13:47:11,51.98918,0.95871,00090,0,0,5,31.1,4.246,inf,51.98923,0.95856,0.0,0,0,0,0,N9898
LORA0: $$KW01,930,13:47:14,51.98915,0.95886,00083,0,0,5,31.1,4.246,inf,51.98923,0.95856,0.0,0,0,0,0,N466D

Appretiate you looking into this.

Cheers
Kev

[daveake]

Still no inf or nan. I'll run XABEN11 again ... Can you send your pisky.txt please. Dave

…

On Wed, 24 Aug 2022 at 15:22, David Akerman ***@***.***> wrote: Inf is expected at the end, when the altitude stops reducing. It's not unusual for this to happen before the code recognises the landing and disables the prediction code. Any inf value prior to landing is wrong. Dave On Wed, 24 Aug 2022, 15:15 Kevin Walton, ***@***.***> wrote: > XABEN11 now finished here, and as per my run last night it worked at the > very end: > > LORA0: $$KW01,927,13:47:06,51.98930,0.95826,00109,0,0,5,31.0,4.246,inf, > nan, nan,0.0,2147483647,0,0,0,N > *E15C LORA0: > $$KW01,928,13:47:09,51.98926,0.95841,00103,39,103,5,31.0,4.246,inf, nan, > nan,0.0,2147483647,0,0,0,N*AE8B > LORA0: > $$KW01,929,13:47:11,51.98918,0.95871,00090,0,0,5,31.1,4.246,inf,51.98923,0.95856,0.0,0,0,0,0,N > *9898 LORA0: > $$KW01,930,13:47:14,51.98915,0.95886,00083,0,0,5,31.1,4.246,inf,51.98923,0.95856,0.0,0,0,0,0,N* > 466D > > Appretiate yuo looking into this. > > Cheers > Kev > > — > Reply to this email directly, view it on GitHub > <#49 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAIGQ5FC3KYSDULOSYOHUSTV2YU7PANCNFSM57KD4LXQ> > . > You are receiving this because you are subscribed to this thread.Message > ID: ***@***.***> >

[KevWal]

Sure - https://github.com/KevWal/pits/blob/develop/boot/pisky.txt

[daveake]

Thanks. Second run with XABEN11 showed no inf or nan. I'll try with your settings next.

…

On Thu, 25 Aug 2022, 08:26 Kevin Walton, ***@***.***> wrote: Sure - https://github.com/KevWal/pits/blob/develop/boot/pisky.txt — Reply to this email directly, view it on GitHub <#49 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIGQ5G5UJSJ3O5U3G3FCW3V24N2FANCNFSM57KD4LXQ> . You are receiving this because you commented.Message ID: ***@***.***>

[daveake]

Try just printing the values used in the CDA calculation and see what you get.

…

On Thu, 25 Aug 2022 at 12:19, Kevin Walton ***@***.***> wrote: This morning I turned all your printf's on inside prediction.c and now I can no longer reproduce it either! I might just add some 'isnormal()' guards around those values for now so it doesn't break the sentance and keep an eye on it.. Cheers Kev — Reply to this email directly, view it on GitHub <#49 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIGQ5AIBY32FXTH5MKEN4LV25JCJANCNFSM57KD4LXQ> . You are receiving this because you commented.Message ID: ***@***.***>

[KevWal]

And I now no longer get it to fail, tried both NMEA files, tried printing or not printing debug stuff, just works every time now, and I cant think of anythnig changed!

I have now added some isnormal() code just in case, dont want it breaking again in flight...

        // Landing Prediction, if enabled and not a seperate sentence
        if (Config.EnableLandingPrediction && (Config.PredictionID[0] == '\0'))
        {
                if (!isnormal(GPS->CDA))
                {
                        printf("ERROR: GPS->CDA not normal!\n");
                        GPS->CDA = 0;
                }
                if (!isnormal(GPS->PredictedLatitude))
                {
                        printf("ERROR: GPS->PredictedLatitude not normal!\n");
                        GPS->PredictedLatitude = 0;
                }
                if (!isnormal(GPS->PredictedLongitude))
                {
                        printf("ERROR: GPS->PredictedLongitude not normal!\n");
                        GPS->PredictedLongitude = 0;
                }
                if (!isnormal(GPS->PredictedLandingSpeed))
                {
                        printf("ERROR: GPS->PredictedLandingSpeed\n");
                        GPS->PredictedLandingSpeed = 0;
                }

Thanks for trying to reproduce. I'll come back here if I see it again!

Cheers
Kev

[PiInTheSky]

OK, I'll protect the CDA code so if there new CDA is silly then it uses the existing value. It's based on altitude changes, and is slugged to reduce the effect of GPS noise and vertical wind gusts. Might be an error around there. Dave

…

On Thu, 25 Aug 2022, 20:46 Kevin Walton, ***@***.***> wrote: And I now no longer get it to fail, tried both NMEA files, tried printing or not printing debug stuff, just works every time now, and I cant think of anythnig changed! I have now added some isnormal() code just in case, dont want it breaking again in flight... // Landing Prediction, if enabled and not a seperate sentence if (Config.EnableLandingPrediction && (Config.PredictionID[0] == '\0')) { if (!isnormal(GPS->CDA)) { printf("ERROR: GPS->CDA not normal!\n"); GPS->CDA = 0; } if (!isnormal(GPS->PredictedLatitude)) { printf("ERROR: GPS->PredictedLatitude not normal!\n"); GPS->PredictedLatitude = 0; } if (!isnormal(GPS->PredictedLongitude)) { printf("ERROR: GPS->PredictedLongitude not normal!\n"); GPS->PredictedLongitude = 0; } if (!isnormal(GPS->PredictedLandingSpeed)) { printf("ERROR: GPS->PredictedLandingSpeed\n"); GPS->PredictedLandingSpeed = 0; } Thanks for trying to reproduce. I'll come back here if I see it again! Cheers Kev — Reply to this email directly, view it on GitHub <#49 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4ERKZNZQXXKGE4VAY4TM3V27ESTANCNFSM57KD4LXQ> . You are receiving this because you commented.Message ID: ***@***.***>