A causal reasoning approach to DSA situational awareness and decision-making

Dynamic Spectrum Access (DSA) systems combine situational awareness development, decision assessment, and spectrum adaptation to provide greater spectrum access to wireless systems. While significant progress has been made in system dynamics and policy conformance reasoning, concern still exists with regards a DSA system's ability to reliably determine the operating conditions for policy compliance in situ. Current methods in literature and recent FCC policies generally develop global operating constraints based on a priori modeling and analysis, which potentially reduces DSA system performance in all cases in order to mitigate risks that occur only in a few cases. Furthermore, the a priori determination of in situ operating constraints is counter to the premise of a “smart” or “cognitive” wireless system. This paper asserts that DSA processes are inherently causal and motivates a causal inference approach to DSA situational awareness and decision-making. The specific method presented is built on Structural Causal Modeling (SCM) and multiattribute decision theory. The approach provides a formal mathematical foundation for developing a core inference structure based on first-order logic while incorporating uncertainty inherent in wireless communications and DSA operations. The proposed approach similarly guides the general awareness and decisionprocessing architecture. A DSA probabilistic SCM development approach based on Multi-Entity Bayesian Network (MEBN) theory is presented along with a worked example demonstrating the potential performance and flexibility gained through in situ probabilistic causal reasoning.