My suggestion (for what it is worth) is to attach small strain gauge-like "feathers" to the wing surface. This would mimic a bird wing where I suppose the bird has some input as to when the feathers move out of their streamlined position, indicating stall or reversed airflow.
Lots of input...