GPS PATH vs. IMU PATH

[john andrews]

Hello All,

I recognized that Google Earth always shows me two paths(GPS and IMU), whether I am in dead reckoning(dead reckoning enabled) mode or gps mode(dead reckoning disabled). Can any body illustrate the difference between GPS and IMU path?

Regards,

John

[Guy-François MESNIL]

Hi John,

Could you give us what kind of data (name of MatrixPilot variables) did you send to google earth, please?

[john andrews]

Hi gfm, 

In the option.h file, i define the SERIAL_OUTPUT_FORMAT as a SERIAL_UDB_EXTRA , with addition of an OpenLog module connected to the serial pins of UDB,

so the UDB will send (to the OpenLog/SD card) a text file that contains the necessary data of navigation like this document  explain, also i convert the .txt file to .kmz(google earth ) file using flight_analyzer  which included in the tools directory of MatrixPilot project . 

Best wishes,

John,     

[william premerlani]

Hi John,

The GPS path is based on the raw data reported by the GPS which will typically be updated at 1, 2 or 4 Hz, and which may have some latency, as much as 1 second. By itself, GPS data does not have enough bandwidth for control and navigation.

The IMU path is computed mainly from accelerometer and gyro signals, is updated internally at 40 Hz, and has very little latency. The IMU path computations incorporate the GPS data to prevent build up in bias due to residual accelerometer offsets. The IMU path is used for control and navigation.

Regards,

Bill

--
--
---
You received this message because you are subscribed to the Google Groups "uavdevboard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uavdevboard...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/uavdevboard/f0c5fa1b-e312-4229-86d9-048101bc1b5f%40googlegroups.com.

[john andrews]

Hi,

Is there anyway I can disable dead reckoning just to do some comparisons. I am happy with it!

John

On Monday, December 30, 2019 at 7:08:07 PM UTC-8, william premerlani wrote:

Hi John,

The GPS path is based on the raw data reported by the GPS which will typically be updated at 1, 2 or 4 Hz, and which may have some latency, as much as 1 second. By itself, GPS data does not have enough bandwidth for control and navigation.

The IMU path is computed mainly from accelerometer and gyro signals, is updated internally at 40 Hz, and has very little latency. The IMU path computations incorporate the GPS data to prevent build up in bias due to residual accelerometer offsets. The IMU path is used for control and navigation.

Regards,

Bill

On Mon, Dec 30, 2019 at 8:20 AM john andrews <j.andr...@gmail.com> wrote:

Hello All,

I recognized that Google Earth always shows me two paths(GPS and IMU), whether I am in dead reckoning(dead reckoning enabled) mode or gps mode(dead reckoning disabled). Can any body illustrate the difference between GPS and IMU path?

Regards,

John

--
--
---
You received this message because you are subscribed to the Google Groups "uavdevboard" group.

To unsubscribe from this group and stop receiving emails from it, send an email to uavde...@googlegroups.com.

[Guy-François MESNIL]

Hi John,

The simpliest way to avoid that the IMU will be corrected by the GPS data is to add this line :

dead_reckon_clock = 0;

just before the test in the module dead_reckon of deadreckoning.c :

void dead_reckon(void)

{

...

dead_reckon_clock = 0; // line added to avoid GPS measurements to be taken into account to estimate the position

if (dead_reckon_clock > 0)

// apply drift adjustments only while valid GPS data is in force.

// This is done with a countdown clock that gets reset each time new data comes in.

{

...

}

As the IMU will be never corrected by GPS data, the position will probably derive quite quickly from the desired flight plan and the navigation will never respect a Waypoint path.

Best wishes,

gfm

[john andrews]

Hello,

Which path(GPS or IMU) should be smoother with no sudden position jumps when shown on Google Earth?

Thanks Bill and GFM for your replies.

Happy New Year!

John

[Peter Hollands]

In Google Earth, the GPS track is working to the nearest centimeter. The IMU track is only accurate to approximately 1 meter. No fractions of meters are plotted for the IMU track.

So the IMU track will appear to jump slightly as it moves each meter without any centimeter changes. The GPS track is smoother with it's centimeter granularity.

Best wishes, Pete

--
--
---
You received this message because you are subscribed to the Google Groups "uavdevboard" group.

To unsubscribe from this group and stop receiving emails from it, send an email to uavdevboard...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/uavdevboard/a087a84c-8477-4de4-91c9-0a1c6376df63%40googlegroups.com.

[Roby]

John,

I think that as Pete said, since the GPS track has centimeter granularity, it should be smoother. But take into account that the IMU path should be more accurate, since it is IMU readings integrated and corrected by the GPS. But Pete or Bill can reply why they chose to display the IMU with meters resolution and the GPS path with centimeter resolution.

Hope this helps

Roby

To unsubscribe from this group and stop receiving emails from it, send an email to uavde...@googlegroups.com.

[john andrews]

Roby,

I am seeing jumps in the IMU path that are greater than one meters so I don't think that is only related to granularity issues as Pete said. I suppose that  is related to time filtering constant not suitable to ground testing..Any idea how to modify it for ground testing?

Best Wishes, John

[Guy-François MESNIL]

Hi John,

I am currently working on the deadreckoning part of MatrixPilot to check if it is convenient, as it is, for my quadcopter flying at low speed and,I expect, precisely.

As the French law obliges us to fly the quad at short range, I am looking for changing the LSB of IMU location from 1 m to 1 cm.

I am using the GP-635-T GPS at 4Hz and UBX protocol.

I can give you some help if you precise your query :

- What GPS do you use and at what rate?

- What do you mean by "ground testing" : are you testing the GPS stationary (inside or outside?) or like a pedestrian?

- Have you some Excel files to show your records?

 

Best wishes,

gfm

[Roby]

Hello Gfm,

Do you think you can achieve 1 cm accuracy just by changing the LSB of the IMU. I suppose that only resolution will improve but accuracy not sure?

Regards,

Roby

[Guy-François MESNIL]

Hi Roby,

The objective given to my work is to guide the quadcopter as precisely as possible, aiming less one meter precision on the trajectory.

To try to achieve that, I need to know the behaviour of all sensors.

Altitude control is complete thanks to the fusion, at low altitude (e.g. <10 m), of the three sensors IMU, Barometer and Lidar.

Horizontal positionning is more complex due to the need of using the GPS, which is not so accurate.

Currently, I try to characterize the GPS in term of positionning (short term (between measurement to measurement) and long term (depending on the GPS own filter selected) and latency.

For this last point, I envisage to compare the GPS to the IMU outputs. So I need the same resolutuion for both, meaning the cm.

I do not hope to guide the drone with a centimeter accuracy.

I just want to compare GPS and IMU for a technical point of view.

Regards,

gfm