Same thought process on conserving battery life.. movement based decision making. The biggest pain with this solution: GPS receiver power consumption. Specifically, tracking current (continuous - hard to get below 3ma) or the amount of current required to wake up and reacquire location data (often 15+ma for 10-30 seconds) . Obviously a "Rhino" can carry a bigger battery. But with many animal tracking applications the battery size vs life becomes the biggest limitation of the product.
There is an "new kid" on the GPS market that proposes doing the complex stuff off line. Can wake up and get a snap shot of the gps signals in a few milliseconds (?). Only records the raw data (no calc of lat/long). Only a few microamps required. drawback: this data "snap shot" is much larger (40K?) than the simple lat/long that would generally be sent out.
For data logging (big memory is cheap) with data that is reviewed later (not transmitted over LoRa) .. makes a lot of sense for long battery life (years).
Pick your poison...short message (more current for gps calc) or long messages (more current for communications channel used)
BTW.. Just finished a new configuration .. on combination of rigid/flex - segmented to allow it to follow curvature of animal's neck (16mm x 100mm total) to fit inside (tubular/hollow) animal collar. While my focus isn't small dogs or cats.. it would likely work on them. Expect to use 1/2 or 2/3 AA battery (added to length of assembly dimensions mentioned) It will continue to bias the weight on collar to keep it positioned correctly (really don't want the radios pointing at the ground). Should add minimal bulk to collar.. and have a battery life (not rechargeable) of over a year (maybe several years).. same basic features mentioned before (LoRa, Bluetooth, GPS, Accel, USB port, etc..)