I've done some work using various radio signals.
In earth environment, Error rate is highly dependent upon signal level, but when errors do occure , they tend to be across multiple consecutive bits.

Known as burst errors, which is why many error-correcting protocols try to spread out the bits protected by each parity bit.

But I'm interested in what @JohnSan's experience was with his system, because it seems somewhat similar to the TS's system, although signal levels may not be comparable. It still gives a first blush at the kind of error rates that the TS should consider.

 

I've not looked into error rates.
My goal was to have the best chance of getting at least some valid data .
Issues being...
Dog collar can move/ rotate on the dogs neck. So aerial position can become unfavourable. Likewise for the GPS receiver also on the collar.

I usually carry the receiver in a trouser pocket, so signal depends on which way I'm facing- towards or away from the dog.

Poor reception is common but in general not a problem.

It only becomes a problem if the dog wanders off and is 'lost'. My control in that situation is i can raise the receiver to a more prominent position- higher the better probably- to get any valid data.
So just one known good reading is enough to attempt finding dog. Hence tx 3 times and try to establish some good data from it..

I then have a means of providing bearing and distance to hopefully relocate him.

Fortunately, I've never had to use it...


Best error correction i think is the reed-soloman encoding, used in writing data on compact discs. But it's probably cpu intensive?

 

Best error correction i think is the reed-soloman encoding, used in writing data on compact discs. But it's probably cpu intensive?

There is no "best" error correction. As with all things, you would first have to define what metric is to be used for "best". Even then, the error coding needs to be well-matched to the characteristics of the error channel it is defending against.

Reed-Solomon, although developed in 1960 to further the math theory behind error-correcting codes, first found use in deep space communications such as Voyager, where retransmission of data wasn't possible. It's actually pretty tame from a computational standpoint, which it had to be if you consider the amount of computational power back in that era, especially for the transmitters. So having your dog's collar encode using it would probably not be much of a burden.