The one that Yuri had built in the post above suffered from issues with airflow being disrupted - too close to the body and flutter of the triangular shaped weather vane part.
How were you thinking of sampling airflow to overcome this?
An accurate airspeed based system would be a massive inmprovement over gps based. Wind correction is no longer needed...
Are you thinking of a pitot like setup in some way?