Car heading in Moving Base Mode

[speedmasters (Evgeny)]

Hello Emlid Community!

I’m new here. We’ve just ordered our first Reach RTK Kit. Our application is autonomous vehicle testing & development. We need a low-cost system for car speed over ground and heading measurement. We need ideally all data in NMEA as we connect Reach to a data acquisition system (kind of logger).

Please share your up to date experience on the following topics:

• Is the speed over ground in Reach output taken from Odometer, specified in ublox NEO M8T Datasheet (see page 14)?

  ublox: “The odometer provides information on traveled ground distance (in meters) using solely the position and Doppler-based velocity of the navigation solution.”

• We plan to use 2 Reach receivers in Moving base mode with a baseline of ca. 2m on the car’s roof. This system should be able to perform quite a precise relative heading measurement. In addition we would also I saw several forum topics with similar issues, but no up-to-date answers from Emlid Team. A flicker of hope are the improvements to the Dec/2017 release in ground speed.

• Reach contains an IMU. How is its data used in the algorithm? Can we expect a True Inertial Heading from Reach? It is normally used in Kalman Filter in RTKLib for Pitch and Roll correction of position, right?

• When I connect in moving base mode my rover via NMEA to an external logger, all the data in NMEA will be the result of moving base RTK correction, right?

• What delays are typical for Reach system in terms of latency and permanent delay to real world events on UART/Serial output?

[speedmasters (Evgeny)]

Dear Emlid Team @egor.fedorov @igor.vereninov @ivan.smirnov,

do you have any answers on this? Sorry for being pushy, but the issues above are quite critical for our applications.

[egor.fedorov]

Hello,

Sorry for the delay. I’ll try answer all of your questions here.

• Reach speed over ground is not taken from the receiver directly, but is calculated inside the RTK engine in the same fashion. You will need the latest dev update for this. It will be rereleased as a stable soon, though

• I’m not sure about where you got the info, but for about a year and a half now, moving-base is not an officially supported mode for Reach units. It is still possible to hack something together, but I can’t make any promises on this.

• IMU readings are not used in the RTK solution. We still plan to implement this, however overall this task is not high on our priority list. Mainly because there are much more efficient techniques to improve navigation results and we decided to focus on them at the moment. The heading will not be “true”, meaning it will be derived from the history of your positions.

• As I said, moving-base is not officially supported. Overall, NMEA will output data according to the mode and solution, not just single.

• Delays very heavily depend on the number of GNSS you are using. In the hardest conditions, the solution should be delayed by about 50 ms. The rest depends on the receiving side.