Optimum welch FFT segment size for duty cycle estimation in spectrum awareness system

In order to realize practical dynamic spectrum access (DSA), implementing spectrum sensing with reasonably low cost is very challenging due to the required detection accuracy and quickness. For this issue, we have introduced an extended DSA, smart spectrum access (SSA), where any useful information in terms of spectrum utilization, such as statistics regarding spectrum utilization, is used not only for accomplishing the requirements of spectrum sensing but also for enhancing the performance of the DSA. In this approach, obtaining the statistics of the spectrum utilization is an important issue for which we have developed a spectrum awareness system prototype. One issue in the spectrum awareness is the Welch FFT segment size design which is used for spectrum analysis. This issue involves a trade-off between the spectrum usage detection accuracy and available frequency resolution. For this issue, we derive an optimum segment size based on analysis and show that the optimum segment size depends on signal-to-measurement bandwidth ratio and duty cycle. The optimum segment size derived by the analysis is validated with numerical simulation and experimental results.