This paper presents a novel architecture and approach for obtaining spectrum availability information in future cognitive radio systems (CRS). CRS can opportunistically access spectrum by identifying unoccupied channels while keeping higher priority systems on the same channel free from harmful interference. Knowledge of the current state of the spectrum use is of utmost importance for CRS operations. There are different techniques for obtaining spectrum availability information including e.g. cognitive pilot channels, databases, spectrum sensing techniques, and combinations thereof. For spectrum sensing there are different algorithms and cooperative combining techniques with different characteristics and capabilities in terms of e.g. performance, complexity, and requirement of a priori information. This paper presents a unified architecture for selecting methods for obtaining spectrum availability information taking into account the operational environment and underlying policies. In addition, a novel low complexity heuristic decision making method is presented for selecting the spectrum sensing technique taking into account different capabilities and requirements while being adaptable to the changing environment.