A comprehensive methodology for opportunistic spectrum access based on residual white space distribution

Abstract

Assuming that Secondary Users (SU) have channel occupancy data (channel idle and busy durations) of Primary Users (PU), this paper presents a methodology to fit primary channel occupancy data to a distribution. Then we show how SUs can opportunistically transmit in the white spaces based on residual white space distribution while keeping probability of interference (with the PUs) below a predefined threshold. We also present analytical formulation for the average throughput of the secondary user and validate it through simulation. Our simulation results show that using the proposed methodology, SU's transmission always remains below the interference threshold and the channel efficiency increases with minimal impact on PU application performance for low values of interference threshold.

Optimal Channel Sensing Strategy for Cognitive Radio Networks With Heavy-Tailed Idle Times

“…Then, we simulate the PU channel occupancy patterns where OFF times are generated with HED distribution given in [8], [12]. When the channel become free from PU, the SU will access the channel using one of the sub-optimal policies.…”

“…From (K + 1) th sensing instant, the SU always choose I * K as the channel sensing interval. In our simulation, we generated the channel occupancy model using two sets of HED parameters given in [8] and [12] (Light traffic -load < 0.1 and Medium traffic -load ∈ [0.3,0.5]).…”

“…The data sets with above behavior have been shown to be well modeled with Hyper-exponential distribution (HED) [10], [11]. Similarly, Sharma et al [12] have simulated 802.11 WLAN clients-server model in OPNET simulator and have observed that the channel idle times can be modeled using HED distribution. The authors of [13] and [14] have proposed an optimal SU sensing / transmission strategy to maximize the throughput of SU with constraint on PU packet collision for generalized as well as hyper-exponential PU idle time distributions.…”

“…The optimal parameters I * 1 and λ * 1e for one-stage sensing interval policy are derived as follows: (i) Evaluate the upper bound on sensing interval, i.e. I using (12).…”

Dynamic Spectrum Access: Current state of the art and future challenges

“…SUs can use various methods to estimate PU traffic characteristics and predict when the PU might appear. A review of the literature finds that there are very few MACs which provide disruption QoS (e.g., [20], [21], [22]). The continued development and verification of protocols that provide disruption QoS may spur wider industry acceptance of spectrum sharing.…”