Dynamic Spectrum Access Decisions. George F. Elmasry

to increase infrastructure resources can use the metrics runtime evaluation to get insight into this need but should use the more comprehensive post‐processing evaluation explained in Chapter 5 to make the ultimate decision of adding infrastructure resources.

      15 15 Inactive here means that there are no end‐user devices utilizing the cell. The cell is always active as a SA.

      16 16 One can see the parallels and differences when comparing this 5G approach to the military communications approach of cooperative distributed and centralized fusion decisions. One can also see that regardless of the problem domain, good DSA design requires considering some form of local, distributed, and centralized spectrum fusion. System designers should be striving to create the appropriate hybrid approach for optimizing spectrum resources use.

      17 17 Local can mean the SA or femtocell has fusion capability, distributed cooperative can mean the different cell types and SAs share spectrum sensing information and make distributed cooperative decisions, and centralized can mean the macrocell is the main performer of spectrum sensing information decision fusion.

      18 18 From the software‐defined network perspective, a centralized spectrum fusion and arbitration in a macrocell means an easier way to roll out new software releases as the software release is targeted to only the macrocell. One can debate if a 5G end‐user device should or shouldn't be software defined but all 5G infrastructure from the femtocell to the largest tower are software‐defined entities where new software releases can upgrade the infrastructure and enhance its capabilities without the need to deploy new hardware.

      19 19 If SAs, femtocells, and picocells are static and their geolocation information is known publically, there is no need to amend spectrum sensing information messages with geolocation information.

      20 20 A block is represented by a frequency band and signal orthogonality at a time slot that is part of a repeated epoch and can be utilized and released dynamically.

      21 21 In practical MIMO antennas, orthogonal signals energy can be detected across different dimensions. This is different from intersymbol interference or co‐site interference.

      22 22 The development of fractional programming and heuristic technique is beyond the scope of this book.

      23 23 Here, the term “optimum” is avoided and the term “feasible” is used instead because the algorithms that can be used may have to be heuristic and have to consider the speed of making decisions and the computational power constraints making them reach a “feasible” decision.