In this talk we consider the problem of designing robust algorithms to detect the presence/absence of wireless microphone signals in uncertain noise plus interference. 
We first show that plain sensing based approaches are non-robust to uncertain narrowband interferers. 
In particular, for energy detector based sensing algorithms, we show that there exists a certain SNR threshold 
below which it is impossible to drive the false alarm probability to zero even when the sensing time approaches infinity.

 We then propose a simple algorithm that combines spectrum sensing with geo-location/database lookup techniques to 
improve the ability to robustly detect the presence/absence of wireless microphone signals. 
Side information in the form of a coarse location measurement, and access to a database that contains information 
about the wireless microphones in that region can significantly improve the performance of spectrum sensing algorithms. 
We give numerical results that show the performance gains from this additional side information.