Primary Signal Detection Performance

4.6 Evaluation Results

the procedure to decide the subcarrier detecting threshold from the subcarrier signals of the received OFDM signal at the master node.

Figure 4.8 shows the simulation results of the false subcarrier estimation prob-ability Pf se and the theoretical result of the signal detection error probability of ASK when the detection threshold is changed. Figure 4.9 shows the ideal subcarrier estimation probability and the variation between the false subcarrier estimation probability and the subcarrier detection threshold. From this figure, the appropriate subcarrier detection threshold must have a high probability for the primary detection. Figure 4.10 shows the results of the probability of detec-tion and false alarm versus the subcarrier detecdetec-tion threshold of an OFDM signal with the fixed primary average SNR at each sensing node as –15 dB. From these figures, when the performance of false subcarrier estimation probability is less than 10⁻⁵, the false alarm probability Pf a becomes stable at 0.1. If the detec-tion probability Pd has a larger value, the subcarrier detection threshold of the OFDM signal, γ_m, has a smaller value. These figures show that a low subcarrier detection threshold influences the probability of false alarm because the master node cannot judge whether the received signal of a subcarrier is signal plus noise or only noise. On the other hand, if the subcarrier detection threshold becomes too large, the subcarrier estimation probability degrades because the master node cannot detect the subcarrier number from the received signal. From these figures, the secondary system can decide the value of the subcarrier detection threshold at the master node as γ_m = 11.5 dB.

Figure 4.11 shows the primary user detection probability versus the mapping parameter at the master node when the primary average SNR of each sensing node is –15 dB. From Fig. 4.11, when the frequency mapping parameter, α, equals to 2.0, the detection probability has a sufficiently large value. Therefore, this subsection evaluates the performance by using α = 2.0. By using these values, the result of the total performance is derived.

sens-4.6 Evaluation Results

0 5 10

10

10

10

Subcarrier Detection Threshold (dB)

False Subcarrier Estimation Probability

Simulation Theory

Figure 4.8: False subcarrier estimation probability,Pf se, versus subcarrier detec-tion threshold,γ_m, at master node. Primary average SNR of each sensing node is –15 dB.

4.6 Evaluation Results

0 5 10 15 20 25 30

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Subcarrier Detsction Threshold (dB)

Subcarrier Estimation Probability

Correct Subcarrier Estimation False Subcarrier Estimation

Figure 4.9: Correct subcarrier estimation probability, P_cse, and false subcarrier estimation probability, Pf se, versus subcarrier detection threshold, γm, at master node. Primary average SNR of each sensing node is –15 dB.

4.6 Evaluation Results

0 10 20 30

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Subcarrier Detection Threshold (dB)

Probability

Detection

False Alarm

Figure 4.10: Primary user detection probability, P_d, versus subcarrier detection threshold, γ_m, at master node. Primary average SNR of each sensing node is –15 dB.

4.6 Evaluation Results

0 1 2 3 4 5

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Mapping Parameter α

Detection Probability

Figure 4.11: Primary user detection probability,P_d, versus mapping parameter, α. Primary average SNR of each sensing node is –15 dB.

4.6 Evaluation Results

ing method, and ideal hard information based cooperative sensing method. In Fig. 4.12, the horizontal axis shows the primary signal SNR at the sensing nodes.

Therefore, this figure can confirm thatPdof the ideal soft information based coop-erative sensing and the proposed method have almost identical performance. On the other hand, the proposed method has a better performance as compared to the ideal hard information based cooperative sensing. Since the proposed method can gather the sensing information from the multiple sensing nodes simultane-ously, the sensing information for primary detection can be gathered by using at least one symbol. In this simulation, it is assumed that the number of symbols for information gathering is one. These results can confirm that the proposed sensing information gathering method can achieve good primary detecting performance with fast information gathering and small power transmission from the sensing node.

4.6.5 Cooperative Sensing Performance with Many Sens-ing Nodes

In the proposed method, sensing information can simultaneously be received from multiple nodes. However, if many nodes transmit the sensing information to the master node simultaneously, the used subcarriers are duplicated among the nodes.

Since these signals are combined on the channel, information on the number of sensing nodes using the same frequency is lost. Hence, this subsection compares the performance of the proposed method with a frequency separated sensing information gathering method.

The frequency separated sensing information gathering method divides the orthogonal subcarrier signals for gathering sensing information into the number of sensing nodes; further, the sensing information is transmitted through the channel using a similar approach as the proposed method. In the system be-ing compared, the channel is separated among the sensbe-ing nodes such that the duplication of the subcarrier does not occur; in this subsection, this method is called as a frequency separated information gathering method. This method is a similar to multiple frequency shift keying (FSK). Here, the sensing information is mapped to the quantized independent subcarrier in each sensing node. Since the

4.6 Evaluation Results

−20 0 −15 −10 −5 0 5

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

: Hard Information : Soft Information

: Proposed

Primary Average SNR of Sensing Nodes(dB)

Detection Probability

Figure 4.12: Detection Probability,P_d, of proposed cooperative sensing. Subcar-rier detection threshold,γ_m, is 11.5 dB.

number of subcarriers of the frequency separated information gathering method is less than that of the proposed method, the quantization bits of the compared method are smaller than those of the proposed method. Therefore, the influ-ence of the quantization error of the frequency separated information gathering method and the subcarrier duplication of the proposed method has a trade–off with the primary signal detection performance.

Figure 4.13 shows the detection probability and false alarm probability of the frequency separated information gathering method and the proposed method.

The horizontal axis shows the number of sensing nodes. The primary average

4.6 Evaluation Results

0 20 40 60 80 100

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Probability

Number of Sensing Nodes : Ideal : Proposed : Compared Detection

False Alarm

Figure 4.13: Primary user detection probability, P_d, versus number of sensing node. Primary average SNR of each sensing node is –15 dB.