Questions about your RTK solution

Some questions about your RTK solution. Specifically the M8P XL version.

  1. Generally what accuracy do you see. I realize this varies by location and weather and KP index. But if you were to go to your testing location what is the normal range of accuracy.

  2. I’ve seen your videos and tried to read as much as possible. I’m interested in how good it is for an inspection scenario where there is as much as 1/2 the sky obstructed to the rover (UAV). This is a typical situation where you want to get close in pictures of a structure. Part of the sky is obstructed. Of course we would get the larger antenna version but I am worried about depending on the accuracy when it cannot be counted on.

  3. My next question is probably more important from a functional point of view. The position accuracy I assume is relative to the location of the base station. So if you have your base at position x and your rover at position y and you record the distance and direction of x relative to y when you go an repeat that position a month later and you precisely put your base station at x and then instruct your rover to move in the direction and distance to y then its position will be within 10 cm (provided locks are all good) of its previous y position.

  4. A bit off topic. It would be very interesting to have 2 gps units on a uav that generate a compass heading. So one of the gps are out on a rod and a heading is given based on their position difference. Compasses are critical to UAVs these days and they are horribly vulnerable to influence. Since you have GPS units (with mags on them) and now processing you might consider this. It is also the only way you can fly UAV in polar regions. In future I might be interested in funding this work.

  5. One last point. Is it possible to setup your RTK solution so that it works without Mission Planner? Say through Tower or Mavproxy?


  1. u-blox announces 2.5cm, we have achived under 5cm without too much difficulty in relatively normal conditions : good antenna + ground plane + open-sky, not that far from buildings.

  2. Very difficult to say, this is 100% dependent on the environment you are going to use the GPS in. Being close to a strcuture means more multipath, which is probably your enemy #1. Dual feed antennas help to attenuate waves with wrong polarization, but this is difficult to model and directly depends on the number of reflections on walls/structures/etc… M8P provide estimators for positioning precision, you can rely on them to check that you are within your accuracy borders.

  3. You are correct, the RTK algorithm provides a relative precision between base and rover (0.025 m + 1 ppm CEP). The overall absolute precision also depends on the positioning error of the base. But if you are locating the base at the exact same position and use the exact same position (not in Survey-in mode but in Fixed mode), then you should find the same measurements.

  4. This is possible and is called “Moving base mode”. A traditional GPS is already capable of generating a “not too bad” heading, as long as it is moving. Moving base uses a modified RTK algorithm and does not need a fixed position for the base. Very useful in order to calculate headings. This functions is already implemented in M8P chips.

  5. You would have to check that these softwares are capable of generating the MAVlink RTCM streams. Or use a different setup to pass differential data directly, not encapsulated in MAVlink.