GPS Referencing for Follow Robot Project

Hello, I'm seeking advice on if this idea could work out theoretically. I do not want to buy the parts to find out it can't work. I plan on having an Arduino Uno installed in a motorized cart along with a adafruit absolute orientation compass, HC-05 bluetooth module, and adafruit ultimate GPS module. I want to make this device follow a remote that is held onto a person (I do not want to use ultrasonics or IR).

If I have an arduino nano, a seperate GPS module, and seperate bluetooth installed in the remote would it be possible to use GPS referencing to get my cart to follow the location of the remote? I plan on using the theory below.

Thanks in advance!

the biggest single problem other than ensuring adequate signal coverage, will be the positioning error introduced by each GPS.

Assuming 10m accuracy, you’ll be looking at some multiple of that with the combined receivers.

So following probably isn’t the right word. Also remember the follower needs knowledge of obstacles or impediments in that 20m wide path it’s taking.

A simple low-power radio beacon and directional antennas would be my starting point. IR could help in the right situations.

GPS works well only outdoors, with a clear view of the sky. At very best you can expect positions to be correct to +/- 2 m, but often with much larger errors.

Does your plan take this into account?

Allround, the Inav would be the simplist to setup, many other advantages as well.

Probably not, if you mean this: GitHub - iNavFlight/inav: INAV: Navigation-enabled flight control software

I plan on implementing an ultrasonic sensor for obstacle avoidance.

Yes it will only be used primarily on a golf course.

Like one with lots of trees on the edge of the fairway, these could affect reception of GPS satellites out on the horizon.

Before going too far with the project, I would hook a GPS up to a logger of some type, then take it for a walk on a golf course.

There should be enough landmarks on the course to allow you to then put the co-ordinates into Google maps to see if the accuracy suits the application.

I just hooked up one of my GPS checker programs and posted the location it produces, in decimal degrees, into Google maps.

I know from the satellite picture exactly where the GPS is located in the roof of my shed, and the GPS reported a position that was 2.2m out.

Bluetooth range is only a nominal 10m.

Has anyone recorded the position computed by two GPS units simultaneously while side by side? That would be interesting.

Around by me the 5th digit of the decimal latitude is a distance of circa 1.1m, and the 5th digit longitude is about 0.7m

From looking at the ouput of the two trackers, right next to each other, whilst often the reported locations are within 1m or 2m of each other, its does regularly jump to a variance of 5m or so, particularly on longitude.

Edit: just seen a variance of 10m on longitude.

Do you see the same randomness in both or do they seem to be independent of each other? To put it another way, if they were both plotted on a map or in u-center, would they move around as a pair (like a couple ballroom dancing) , or would each be off on its own random path?

Overall, long term, then the position error might well move around N,S,E,W, never plotted it.

However short term, the location variance or distance between two GPS can still vary by 5m or so, even when the GPS fixes are in the same second.