bGatti recently made a post that got me thinking: given the plethora of information captured during logging manual flights, could one run a post-analysis (a bunch of linear regressions?) to infer the in-flight characteristics of a plane. Then the optimized PID values could be computed offline and we could also do better HIL simulations. Has anyone explored this? It would greatly lessen the need for an airspeed sensor if we could make an educated guess on the airspeed given orientation, throttle, control surface positions and GPS velocity. Thoughts?
I may have to bother my aeronautical engineer brother - I'm just an ECE :)
Replies
Has it been done in this community - not to my knowledge.
There seems to be a lot of misconceptions about the airspeed sensor. For most airframes and mission types it is not needed with APM. If you want to fly an EasyStar or a Funjet or whatever, and you are not interested in precision landings you probably do not need an airspeed sensor. If you do need an airspeed sensor then the reasons you need it preclude the use of a model for estimation of airspeed. A model cannot tell you when you are being hit by a gust of wind from behind while flying slowly on short final....