Policy Specification and Enforcement For
                    Spectrum-Agile Radios

                                      Grit Denker

           SRI International, 333 Ravenswood Ave, Menlo Park, California 94025
                                   Grit.Denker@sri.com

     Because of the centralized, static nature of current spectrum allotment policy, wire-
less communication is confronting two significant problems: spectrum scarcity and de-
ployment delays. Current systems are significantly impacted by the lack of access to
unused spectrum, and existing spectrum management procedures are too inflexible to
react to dynamic operational needs. The policies are static and policy changes are very
labor intensive.
     Existing policies assume 100% spectrum use; however, studies have shown that
most assigned spectrum is unused most of the time. This motivated the goals of DARPA’s
neXt Generation (XG) Communications Program, which envisions opportunistic spec-
trum access. XG provides technologies for automatic, dynamic, and opportunistic ac-
cess to unused spectrum. This requires that radios sense and opportunistically adapt to
local RF environments and application needs.
     In addition, radios must act according to regulatory rules. Spectrum use policies are
authored in more than 200 countries and are verified by each host nation. Policies are
customizable to location, user, time, frequency and many other parameters, and they
change over time. The large number of operating dimensions to be considered makes
it difficult to find a solution for the optimal use of radio spectrum. In XG this problem
is solved by using platform-independent Policy Controls to regulate spectrum access in
changing regulatory environments.
     We developed an expressive and extensible policy language for describing policies
that meet the needs of a wide variety of spectrum regulation bodies, including reuse
of ontological concepts, and for supporting efficient reasoning. We also implemented
a policy-conformance algorithm that checks the compliance of candidate transmissions
using efficient, state-of-the-art reasoning technology. The design of the algorithm is
independent of, yet easy to integrate into, any radio design, to encourage competitive,
best-of-breed XG radio development.
     We will report on our experience in designing the Cognitive Radio Language (CoRaL).
CoRaL is a declarative language based on a typed version of classical first-order logic.
We will also summarize the results of our evaluation of various logical formalisms with
respect to their appropriateness for the XG domain. Finally, we will illustrate the main
concepts of the XG Policy Reasoner.

Acknowledgments.
This research was supported by DARPA’s neXt Generation (XG) Communications Pro-
gram under Contract Number FA8750-05-C-0230. (Approved for Public Release, Dis-
tribution Unlimited)