Smarter adaptive smart interval for position updates

[RichieFrame]

I have spent the last few hours (more than a few) writing a 'smarter' smart interval, more of an adaptive smart interval that targets 3 intervals, a stationary max, a minimum, and a walk speed... for example:

360s max, 60s walk, 30s minimum, 12m = 201s, 24m = 146s, 200m = 34s
600s max, 90s walk, 45s minimum
1200s max, 150s walk, 90s minimum

The walk speed is based on a 1.4m/s speed, if you travel that fast the interval will adapt to the walk target, if you go really fast, like car driving speed, it will hit the minimum, and when stationary the maximum, and anything in between will adapt to the defined curve, I put some of the calculated values in the 360s example.

I will be testing this maybe tonight or tomorrow, it LOOKS like it would be useful, but need to see it in practice to understand if my algorithm (weakly adapted from an exponential regression curve) is too complex or not complex enough, and if the desired intervals are reached with an accurate GPS source, as this could drain a battery way too fast when not accurate.

Would this be of use to anyone other than me?

[fifieldt]

Yup :) Prerequisite work here: meshtastic/protobufs#749

[RichieFrame]

Well that is interesting, an adaptive interval can absolutely be applied to a profile with a simple profile + frequency (slow/med/fast) selection, I build 9 frequency selections but they all have timings selected for walking speed, I could build some for vehicle speed (25m/s) or bicycle/scooter speed (8m/s)

Max intervals may still need to be configurable to prevent stale data for a use case, and min to prevent excessive updates for a given device's battery size, my use case is pedestrian that MAY be moving fast or even getting in a vehicle, and the adaptive curve gives a good update interval for any speed... in theory, I will see soon

[nasiralamreeki]

Some improvements in this area would be great especially if they reduced power consumption