Originally Posted by
brawlo
The satellites are what they are, it’s the calculation hardware in the unit that gives accurate positioning, particularly the clock. Atomic clocks small enough to fit inside a phone or even a Garmin have been around for years now. Those would enable accuracy to centimetres rather than the standard metres, but the consumer isn’t willing to pay the extra for that level of accuracy. A speed sensor is a whole lot cheaper
You don't need an atomic clock to do it - the velocity measurement is just a doppler shift measurement, and it's one that every GPS (including cheap ones) has to do in order to lock on the satellite signals. Consumer-priced GPS units could do it if people wanted it - there are ~20 year old publications of people doing ~cm/s accuracy with cheap consumer grade GPS units. A magnet and a frequency counter are a lot cheaper, but given the large numbers of expensive GPS units I see on bikes, it wouldn't be unreasonable to expect it from GPS units. I really just wanted to address the suggestion that GPS speed is intrinsically inaccurate (that Carleton implied) - even consumer grade GPS *can* give very accurate speed, they just aren't sold with that feature for cyclists.