Problem to get Ready status when using IM19 with GNSS Flex pHAT ZED-X20P

Sensors IMU & Motion
1

Hi all,

I’m having problem to get Ready status after setting up ZED-X20P(using u-center configuration) and IM19 as follows:

PPS was set with pulse period of 1000ms and pulse length 100ms

CFG_UART1OUTPROT CFG_UART1OUTPROT_NMEA 1
CFG_MSGOUT_NMEA CFG_MSGOUT_NMEA_ID_GGA_UART1 1
CFG_MSGOUT_NMEA CFG_MSGOUT_NMEA_ID_GLL_UART1 0
CFG_MSGOUT_NMEA CFG_MSGOUT_NMEA_ID_GSA_UART1 0
CFG_MSGOUT_NMEA CFG_MSGOUT_NMEA_ID_GST_UART1 1
CFG_MSGOUT_NMEA CFG_MSGOUT_NMEA_ID_GSV_UART1 0
CFG_MSGOUT_NMEA CFG_MSGOUT_NMEA_ID_RMC_UART1 1
CFG_MSGOUT_NMEA CFG_MSGOUT_NMEA_ID_VTG_UART1 0
CFG_NMEA CFG_NMEA_HIGHPREC 1
CFG_RATE CFG_RATE_MEAS 200

AT+MAG_AUTO_SAVE=DISABLE
AT+AHRS=DISABLE
AT+NAVI_OUTPUT=UART1,ON
AT+LEVER_ARM=0.007,-0.035,-0.025
AT+CLUB_VECTOR=0.00,0.00,0.00
AT+GNSS_CARD=OEM
AT+HIGH_RATE=ENABLE
AT+SAVE_ALL

Here is some outputs from the IMU:
$GPFMI,111043.00,0.00000000,0.00000000,0.000,3.13909,0.00232,9.99999,0.452,12,00,2,001C000110
$GPFMI,111043.20,0.00000000,0.00000000,0.000,3.13911,0.00232,9.99999,0.452,12,00,2,001C00011B
$GPFMI,111043.40,0.00000000,0.00000000,0.000,3.13912,0.00229,9.99999,0.452,12,00,2,001C000114
$GPFMI,111043.60,0.00000000,0.00000000,0.000,3.13914,0.00226,9.99999,0.452,12,00,2,001C00011F
$GPFMI,111043.80,0.00000000,0.00000000,0.000,3.13913,0.00228,9.99999,0.452,12,00,2,001C0001*18

I am not entirely sure how the lever arm should be set because the dev board is on the table for me to test out the accuracy of the IMU when fused with GNSS RTK positioning. I plan to use this IMU for agriculture machine automation. Did I misconfigure or miss any critical step here?

Thank you

2

Hi @cgi27 ,

There is a lot of useful information in the post linked below. But I’m guessing you may have already seen it?

Sadly, you can’t test this on the table. The IM19 and GNSS antenna need to be physically coupled together, the antenna needs to be outdoors, and you need RTK Fix. Only when you have all three can you start the IM19 calibration.

Sorry! But that’s just the way it is. ‘Faking it’ with the antenna outdoors and the IM19 in the warm and dry doesn’t work…

Also, please check you are using the correct WORK_MODE. I’m not sure which one is going to be best for your application. Probably 72?

image
image588×263 39.8 KB

I hope this helps,
Paul

3

You are enabling GGA, GST, and RMC messages, which are settings for

AT+GNSS_CARD=OEM

You should download the IM19 Integration Guide if you haven’t already done so. I believe your choice here should be OEM.

4

Thank you for pointing that out. Oh sorry that was set to OEM but it does not change the STATUS to ready

5

It will either be 13 or 72. Have you tested the accuracy of the NAVI data esp. roll, pitch, yaw with and without RTK fix after the calibration? Because there will be times that RTK fix won’t be available for my application

6

The IM19 uses Kalman filtering that requires a time history of measurements from multiple sensors. There’s a lot of statistical theory involved. I don’t claim to be an expert, but I have studied it in the past.

Calibration requires input from all sensors to complete initialization. Nearly continuous use is needed to maintain ready status. Kalman filters work by using statistical methods, starting from prior knowledge, making predictions based on a physical model, taking new measurements, updating and reporting internal state, projecting the state of the model forward in time, and predicting the new state. The process repeats. If measurements and predictions diverge too much, calibration is “lost” as can be seen by a change in the reported status.

I have not checked accuracy, but I have observed that pitch and roll appear to be correct even in the Finit state. Yaw needs RTK fix and “walking” or moving sufficient distance to determine orientation. Once it is completely calibrated, I have noticed that the status will eventually degrade if there is no motion.

I have tried two of the Tilt modes. I have not tried Vehicle nor Heading modes. I suspect you will need RTK fix and sufficient motion for calibration. Feyman claims DR continues to be accurate for 60 second GNSS outage.

1 Like
7

Thank you for the explanation. It makes a lot of sense now that I need to be outdoor to test it out.