Dynamic Spectrum Access Decisions. George F. Elmasry
target="_blank" rel="nofollow" href="#fb3_img_img_2c1a0656-2150-5a74-ae30-82357c376e17.png" alt="equation"/>
The selection of the feasible set considers many factors, as explained earlier, including minimizing SIR, maximizing date rate, minimizing power dissemination, and using sensing information to avoid interference with other systems.
Now let us look at a global metric a centralized DSM arbitrator may consider to achieve this multifaceted global energy efficiency, EEG which can be expressed as:
(6.12)
Notice that in Equation (6.12) the nominator is the global data rate while the denominator is the summation of assigned power. The search for an assigned power set to all transmit/receive pairs can be weighed based on maximizing this global metric (among other metrics). Notice that EEG does not consider the device circuits' efficiency. The device circuits' efficiency is the ratio of dissipated power as spectrum emission to the assigned power vector. Also this optimization does not consider the presence of AWGN.
EEG is one of many metrics the centralized arbitrator can use. Note that the more metrics one attempts to include in a spectrum fusion technique, the more complicated the technique can become. One has to look for the most effective modeling of the problem at hand and consider the most critical metrics that can help with the development of a practical cognitive spectrum fusion technique that can be evaluated in real time and their post processing can bring about meaningful policy and configuration parameters changes.
6.6 The Role of the Cell and End‐User Devices in 5G DSM
The optimization problem described in the previous section may require the collection of information from the end‐user device in order to ascertain if the central arbitrator resource blocks and power needs to be adjusted. Recall that Equation (6.12) did not consider the presence of AWGN or the relationship between allocated power and dissipated spectrum, which requires getting some performance metrics from the end‐user device to adjust the allocation of resource blocks and power. The end‐user device's lower protocol stack layer can collect different measurements such as block error rate, latency, jitter, and channel quality. This data can be utilized by the upper protocol stack layer algorithms for real‐time adaptation of traffic load pushed to the lower stack layers and can also be abstracted and forwarded to the centralized arbitrator to fine‐tune the global resource allocation modules.
It is important to note that the end‐user devices and the cell access points also have some functions that help reduce the need for the central arbitrator to continually readjust resource allocations. These functions can include the following:
A cycle adaptation function that is responsible for managing coexistence with other transmit/receiver pairs. This function adopts the time‐domain utilization pattern.
A listen‐before‐talk (LBT) function, which is tied to the coexistence function through detecting the state of the frequency resource prior to the data transmission. This function manages energy detection, preamble detection, and duration. This function can rely on tunable thresholds to compare the sensing results with.
A frame format function, which is responsible for adapting the MAC frame according to the different active bearers (e.g., uplink versus downlink resource ratio, transmission time interval [TTI], or slot duration modification). This function can also react to the changes in the channel quality, enabling better coexistence without the need for intervention from higher hierarchy entity (e.g., cell).
A contention coordination function, which manages any perceived contention on the utilized frequency bands. This function can adapt the random access schemes and block resources on licensed bands or can use a tunable contention access algorithms on shared spectrum bands.
A multiple access (MA) function, which is responsible for the configuration and adaptation of the use of spectrum resources utilizing orthogonality and nonorthogonal multiple access schemes. Within the cell, this function can manage different active end users given the end‐user locations and QoS requirements.
A sensing function that coordinates between the different sensing mechanisms. This function can create sequential sensing patterns on the different bands while using configurable parameters such as sensing duration, minimum signal detection level, and sampling rate to create effective spectrum sensing information.
Within the cell to end‐user radio access technology (RAT), the above functions make short‐term decisions leaving longer‐term decisions to the centralized function.
Notice that there are other aspects of DSM that have to be considered by the developed DSA technique, including the following:
Traffic demand can be high in one area and low in another area. The centralized arbitrator allocation of spectrum resource blocks to the different RATs in different areas of the network it is managing can take into consideration traffic demand over time. The resource blocks (or the network infrastructure) are essentially a shared service between the different RATs.
The small cell has limitations in handling traffic loads at a required QoS. Throwing more resource blocks to a small cell may be a waste of resources given the small cell limitations.
Ultra‐dense urban deployments can force the DSM central arbitrator to create layered architecture of spectrum sharing. The DSM technique would need to consider the tradeoff between spectrum and network density to optimize the network spectral efficiency of multiple RATs sharing a spectrum resources pool.
Other aspects include the goals of the service provider. Some service providers may market guaranteed QoS for higher prices to attract high paying customers while others may market lower prices with less QoS guarantees to create a mass market. These revenue‐focused aspects will drive DSM implementation in the 5G infrastructure.
6.7 Concluding Remarks
Although the concept of local, distributed, and centralized spectrum sensing decision exists with 5G, one can see how it can differ from the military network examples used in Part 1 of this book. One can also see how the hybrid DSA design can be different in some aspects while others stay the same. Aspects that stay the same include sensing in geographically distributed locations, creating the framework for hierarchical fusions, and sharing of spectrum sensing information in a combined distributed and centralized manner when possible. The aspect of making DSA a cloud service with well‐studied metrics that can be used to evaluate DSA services in real time and in post processing is also common between the many DSA cases.
5G gives the service provider many flexibilities in managing the infrastructure. One can expect a diverse set of DSM techniques to be present in different services providers' networks. However, there is a common theme in these optimization approaches where metrics are used and cognitive techniques are used to approach optimality of spectrum assignment. Relying on a good model for spectrum sharing and using expressive metrics to measure performance is always needed. Also, all service providers are likely to approach DSM as an IaaS set of cloud services, as explained in Chapter 5. One can expect that any service provider approach will have to be hybrid in one way or another and will have to morph based on deployment constraints and real‐time measurements.
In some aspects, DSA for military networks may seem easier than commercial 5G, but the expected use of commercial technologies within military communications systems can complicate military network DSA decisions. The push for military networks to use unlicensed spectrum is another factor that can make military system DSA techniques more challenging.
This chapter presented a sample of DSM techniques for 5G. There is a wealth of literature sources for 5G DSM that further explain the developing of mathematical models for DSM, explain