VIC3DA Dead reckoning not working

[yclin]

I have some problem with DR mode.

I can't recieve any $PSTMDR2 or $PSTMDRCAL messages, so the status of DR remains unknown.

But I checked the CDB 228-230 with $PSTMGETPAR, the following are the response.

========

$$PSTMSETPAR,1228,0x00000000*5A
$$PSTMSETPAR,1229,0x6cc22000*5D
$$PSTMSETPAR,1230,0x00000000*53

==========

In CDB 229, bit 60-63 is 6, meaning that the $PSTMDR and $PSTMDRCONFID are enabled, but I can only see $PSTMDRCONFID. There are no $PSTMDR message there. Are there any problem with my configurations?

Also, I had enabled these messages in CDB 229, meaning that these message should be update in 10 Hz. But it still be updated in 1 Hz, as following.

==========

$$PSTMDRCONFID,0.00,0.00,0.0000,0.0000,0.00,0.0000,0.000000,0.000000,0.0000,0.0,0.0,-0.0,0.0*1A
$$PSTMDRPVA,162640.000,2945690305,37.5517719,14.2984835,730.1,0.00,0.00,0.00,0.00,0.00,0.00*57
$$PSTMDRCONFID,0.00,0.00,0.0000,0.0000,0.00,0.0000,0.000000,0.000000,0.0000,0.0,0.0,-0.0,0.0*1A
$$PSTMDRPVA,162641.000,2946713305,37.5517719,14.2984835,730.1,0.00,0.00,0.00,0.00,0.00,0.00*5F
$$PSTMDRCONFID,0.00,0.00,0.0000,0.0000,0.00,0.0000,0.000000,0.000000,0.0000,0.0,0.0,-0.0,0.0*1A
$$PSTMDRPVA,162642.000,2947736306,37.5517719,14.2984835,730.1,0.00,0.00,0.00,0.00,0.00,0.00*59
$$PSTMDRCONFID,0.00,0.00,0.0000,0.0000,0.00,0.0000,0.000000,0.000000,0.0000,0.0,0.0,-0.0,0.0*1A
$$PSTMDRPVA,162643.000,2948758283,37.5517719,14.2984835,730.1,0.00,0.00,0.00,0.00,0.00,0.00*53

=========

Are there some configurations needs to be changed? Thank you.

[GalaxyQuest]

Hi,

CDB-228 contains all the $PSTMDR messages, so having values all 0s explains why you are not seeing any PSTMDR messages.

Can you please refer to $PSTMSETPAR command on how to set the bits. It looks like you might have overwritten the CDB registers.

[yclin]

Hi @GalaxyQuest 

Thank you for your response. I had understood that these DR messages report at 1 Hz, so these message should be set in CDB 228 (that include 1 Hz messages)

Also, I can recieve these DR messages after a factory reset. Just a note if someone needed.

CDB 228: 0x6cc22000   CDB 229: 0x0 CDB 230: 0x0

CDB 600: 80A00116 (there is a conflict that the manual says the bit 31 should be 0, but the default setting is 1)

=================

My new question is how to improve the DR performence?

I followed the instructions written in AN5678, and I got "system ready" flag.

I tested DR in a 2 km tunnel. But when I leave the tunnel, the position shift 500 m when the GPS was available.

Are there some values that I can know the accuracy of the DR? like the error in PSTMDR2?

Thank you.

[GalaxyQuest]

Hi,

Couple of things to check.

I am assuming you are using DRUM mode -- i.e. no connections to vehicle, is that correct?

If so, in PSTMDR2 - you should see SYSTEM_READY along with status of 0xCC should be seen in IMU cal.

0xCC means accelerometer calibration and gyro calibration is completed.

Secondly,

Could you please ensure that when doing field testing, the device is firmly placed on the vehicle frame and not placed on a seat where it can move around?

Regarding DR messages.

In PSTMDR2 we report the cross track and along track error which is a good reference.

In addition, we also report PSTMDRPVA - which is DR estimated position and PSTMDRGPS - which is pure GNSS.

[GalaxyQuest]

Hi,

You can also monitor PSTMDREPE, PSTMDRPVASD and PSTMDRSVF.

You can enable this by sending the following command.

$PSTMSETPAR,1228,0x200,1      //enable PSTMDRSVF
$PSTMSETPAR,1228,0x2000000,1  //enable PSTMDREPE and PSTMDRPVASD
$PSTMSAVEPAR                  //save
$PSTMSRR                      // reset for setting to take effect

To handle the higher volume of messages, I would recommend increasing baud-rate

$PSTMSETPAR,1102,0xC
$PSTMSAVEPAR
$PSTMSRR

Close and reopen Teseo Suite.

[yclin]

Hi @GalaxyQuest 

Sorry for late response. I have done some test.

I use DRUM mode, and I checked PSTMDR2 with "CC" flag. So the DR mode was ready.

Sometimes I got good result (<50 m error between DR last estimate position and GPS position after driving through a tunnel).

But sometimes the speed estimated by DR would suddenly drop to 0 after I entered the tunnel (and the actual speed was around 80-90 km/h). The flag from PSTMDR2 was still "CC". Seems the DR mode was still working, but it give the wrong data. 

Speed data and X axis acc data are the following: (red box is the data when I entered the tunnel)

I had placed the VIC3DA on a heavy steel plate and tight it with seat belt. It may not have relative movement.

Are there someting I misunderstood?

Thank you.

[GalaxyQuest]

Hi,

Can you please provide me what version of firmware you are using in VIC3DA?

Can you please provide a log - if this log includes when the speed in tunnel is normal and when speed in tunnel drops to 0, that would be helpful.

Could you please let me know what $PSTMDRPVA reports when speed drops to 0?

---

[yclin]

Hi @GalaxyQuest 

Sorry. Previously, the data was modified to avoid violating confidentiality agreements with my work partners. Now that the matter has been clarified, please consider the following data as the accurate reference.

We have collabrate with a railway company. They want to get the accurate position of trains by GPS&DR even if the train is in the tunnel. So we chose VIC3DA to do that.

Because the route of the railway doesn't meet the requirement of AN5678, we trained DR with another car and put VIC3DA on a steel plate with horizontal adjustment. And we moved the steel plate on the train, did horizontal adjustment, trying to simulate the situation in the car.

Although the "CC" flag in PSTMDR2 is still there when the train entered the tunnel, but DR seems not working. Is that we need to adjust the setup more accurate? Or are there some better way to get accurate position in the tunnel, or using DR easily?

Thank you.

22.05.25_16.08.21_NMEA_Device_D0 log file is the data collected on a car and had well-trained DR.

26.05.25_13.12.24_NMEA_Device_D0 log file is the data collected on the train. 

version of firmware: 4.6.8.11

[GalaxyQuest]

Hi,

I was looking at the data that you provided me.

I see that the log with 26,05.25 date tag, the result was done in good sky visibility and you got good results as expected.

The log with 22.05.25 date tag, I see the vehicle went under a tunnel and here again I see good results. I was looking at the speed data and did not observe any drop-outs to 0. I think for this run, the data looked good, the vehicle maintained its position in the tunnel.

I am attaching the results I got from analyzing the log.

Since this looks like a commercial project, are you working directly with ST sales or distributor sales? You can private message me if you would like with the information and I can try to connect you with a local team.

Thank you.

[GalaxyQuest]

In addition to my above message, could you please let me know how the antenna is placed? 

The reason I am saying is during the initial part of the run when you are in open sky, the signal levels are not that great. In open sky, you would except CN0 of the signals to be ~45 db-Hz while most of your signals are in 30s and 20s.