@Brittany Thompson good question and honestly a really practical one — canopy is where GNSS gear lives or dies for trail applications.
The honest answer is that under thick deciduous canopy you're probably looking at 3-6 meters on average, with occasional spikes worse than that during really dense patches or terrain features that cut your sky view down even further. The 1.5m spec is open-sky CEP, so yeah, that's not what you're getting in the woods — but 3-6m is actually still workable for trail mapping if you're not trying to distinguish between two parallel trails 5 meters apart. Your GPX track will look like a trail, not a random walk. The 18Hz rate genuinely does help here because you're averaging more frequent noisy fixes rather than fewer noisy fixes — the noise tends to be somewhat random so the higher sample rate gives your post-processing more to work with.
The IMU dead reckoning idea is solid but you're right that it's a software fusion problem, not something that happens natively on the hardware. The sensors module gives you the raw accelerometer and gyro data, but fusing that with GNSS into something like a Kalman filter is work you'd have to implement yourself. It's not trivial, but it's also not a PhD project — there are open implementations you could adapt. The bigger practical issue is that IMU dead reckoning drifts pretty fast on foot (unlike a vehicle with a wheel encoder to anchor it), so it's more useful for bridging a 2-3 second signal dropout than compensating for sustained canopy attenuation over a 10 minute stretch.
One thing that helps more than people expect in dense canopy is just raising the antenna position — even a few extra centimeters of clearance on a chest or shoulder mount vs. waist level changes your sky view meaningfully. Worth experimenting with before going deep on sensor fusion.
