Temel Kavramlar
The authors propose an expectation maximization and Viterbi algorithm based scheme to estimate the states of a dynamic primary user and use this information to improve the detection performance of a modified weighted sequential energy detector in a distributed cooperative spectrum sensing network.
Özet
The content discusses a distributed cooperative spectrum sensing (DCSS) scheme for detecting a dynamic primary user (PU) in a cognitive radio network. The authors first review the conventional energy detector (ED) and the weighted sequential energy detector (WSED) approaches, and then propose a modified WSED (mWSED) algorithm.
The key highlights are:
- mWSED aggregates only the energy samples that correspond to the present state of the PU, unlike WSED which aggregates all the present and past samples.
- Since the PU states are unknown in practice, the authors develop a joint expectation maximization (EM) and Viterbi algorithm to estimate the PU states from the collected energy samples.
- The estimated states are then used in mWSED to compute its test statistic, resulting in the EM-mWSED algorithm.
- Simulation results show that both EM-Viterbi and EM-mWSED outperform the conventional ED and WSED approaches in detecting the dynamic PU, especially by increasing the network connectivity or the SNR/number of samples.
İstatistikler
The number of secondary users (N) in the network is 10, 20, or 60.
The network connectivity (c) is 0.2 or 0.5.
The number of samples per energy statistic (L) is 12 or 36.
The SNR values considered are -5 dB, -3 dB, or 0 dB.
The primary user follows a two-state Markov chain model with transition probabilities α = β = 0.1.
Alıntılar
"The authors propose an expectation maximization and Viterbi algorithm based scheme to estimate the states of a dynamic primary user and use this information to improve the detection performance of a modified weighted sequential energy detector in a distributed cooperative spectrum sensing network."
"Simulation results show that both EM-Viterbi and EM-mWSED outperform the conventional ED and WSED approaches in detecting the dynamic PU, especially by increasing the network connectivity or the SNR/number of samples."