Performance Evaluation

In this simulation, the number of WLAN users is increased from 8 to 69 and the number of WCDMA users is increased from 20 to 100. we define the offered WALN traffic load as the ratio of the total average arrival rate of all WLAN users over the system maximum transmission rate which is 11 Mbps in 802.11b. Note that the average arrival rate of WLAN user is equal to 145 kbps. Thus, the offered WLAN traffic load varies from 0.1 to 0.9 as the number of WLAN users varies from 8 to 69.

In the performance evaluation, we compare the systems consisting of WLAN and WCDMA to the systems consisting of WLAN, WCDMA and CR. The duty cycle values are chosen as follow:

Duty cycle values:

The basic unit frame (BUF) of each duty cycle values is 100 frames. Another BUF of duty cycle values is 50 frames. The resulting duty cycle values are shown in Table 4.5. We also compare the performance between two different BUFs.

Table 4.5: duty cycle values

BUF = 100 frames Duty cycle values:

WLAN

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Fig. 4.3: The individual CR throughput

Fig. 4.3 illustrates the individual CR throughput in WLAN spectrum and WCDMA spectrum. It can be found that CR systems obtain large throughput in WLAN spectrum when the offered WLAN traffic load is light. With increasing offered WLAN traffic load, CR systems operating in WLAN spectrum achieves a smaller throughput and almost zero at high offered WLAN traffic load. The reason is that CR systems using CR-based DSA algorithm in WLAN spectrum will adaptively choose proper duty cycle values according to the estimated channel load. On the other hand, CR systems operating in WCDMA spectrum obtain no throughput at light offered WLAN traffic load. When offered WLAN traffic load increase, the throughput of CR systems in WCDMA spectrum increases slowly and finally becomes a constant value. The reason is that CR-based DSA algorithm using frequency bands switch shifts the operational spectrum from WLAN spectrum to

WCDMA spectrum with a higher probability when the offered WLAN traffic load is larger. However, the resource in WCDMA spectrum is limited by the primary users, WCDMA users. Only the residual power of WCDMA B.S can be shared to CR systems. Therefore, the maximal throughput of CR systems in WCDMA spectrum keeps a constant value. In addition, the performance of CR with 50 BUF is better than that with 100 BUF at middle offered WLAN traffic load shown in Fig. 4.3. The reason is that when the BUF becomes larger, the active periods and quiet periods of CR systems become larger. The larger active periods cause long packet delay and serious collision problems to WLAN systems. During the quiet periods, the larger active periods cause higher estimated WLAN traffic load than the smaller active periods. Therefore, CR systems have larger opportunity to access channels with larger duty cycle value when BUF becomes smaller. Hence, at middle offered WLAN traffic load, CR systems with smaller BUF perform better performance in throughput.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Fig. 4.4: The Throughput of WLAN Systems

Fig. 4.4 shows the throughput of WLAN systems. The resulting throughput of a WLAN system without CR is shown in Fig. 4.4. In WLAN systems, the probability of successful transmission increases when offered WLAN traffic load increases at light and middle traffic load. Hence, the throughput of WLAN systems increase with offered WLAN traffic load. This phenomenon is expected since current channel resource is larger than current users’ needs. However, at heavy traffic load, the probability of successful transmission decreases and collision increases when offered WLAN traffic load increases. Therefore, the resulting throughput of WLAN systems decrease with offered WLAN traffic load. The reason is that the offered WLAN traffic load increases with the number of WLAN users. More uses in a WLAN system, more contention it causes. With increasing contention, the probability of successful transmission decreases and the probability of collision

increases. On the hand, the resulting throughput of a WLAN system with CR is also shown in Fig. 4.4. It is obviously that resulting throughput of a WLAN system with CR is as much the same as that without CR. The reason is that CR systems with CR-based DSA algorithm do not suppress the resource allocation of WLAN systems.

CR-based DSA algorithm in WLAN spectrum adopts a duty cycle value to decide the active periods and quiet periods of CR systems. The quiet periods protect WLAN users’ rights to transmit data.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Fig. 4.5: The Proportion of idle time in WLAN spectrum

Fig. 4.5 shows the proportion of idle time in WLAN spectrum. The proportion of idle time in WLAN spectrum is the total idle time in WLAN spectrum over the total simulation time. It is clearly that CR with CR-based DSA algorithm improves the idle time in WLAN spectrum at low and middle offered WLAN traffic load. The phenomenon is expected since CR-based algorithm adopts a larger duty cycle value

to extend active periods and shorten quiet periods at low and middle offered WLAN traffic load. As to the residual idle time at low and middle traffic load, CR-based DSA algorithm enforces CR systems to keep quiet for such periods of frames. The reason is that the estimated WLAN traffic load is heavy at high traffic load.

Therefore, CR systems try to release more channel resource to WLAN systems, resulting in the residual idle time large at low and middle offered WLAN traffic load.

In addition, the performance of CR systems with shorter BUF is better than that with longer BUF at middle offered WLAN traffic load shown in Fig. 4.5. The reason is that when the BUF becomes larger, the active periods and quiet periods of CR systems become larger. CR systems with larger BUF will occupy the channel for a long time. The larger active periods cause long packet delay and serious collision problems to WLAN systems. During the quiet periods, the larger active periods cause higher estimated WLAN traffic load than the smaller active periods. Therefore, CR systems have larger opportunity to access channel with larger duty cycle value when BUF becomes smaller. Hence, at middle offered WLAN traffic load, CR systems with smaller BUF perform better performance in idle time reduction.

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

Fig. 4.6: Average packet delay of WLAN systems

Fig. 4.6 shows the average packet delay of the WLAN system. When the WLAN system operates with the CR system, the average packet delay of the WLAN system decreases when the offered WLAN traffic load is smaller than 0.3. The reason is that the CR system measures the current WLAN traffic load is small. Therefore, the CR system with CR-based DSA algorithm chooses the larger duty cycle value to extend the active periods. When the offered WLAN traffic load increases, the CR system measures the current WLAN traffic load is large. Then, the CR system chooses the smaller duty cycle value to shorten the active periods and extends the quiet period.

When the offered WLAN traffic load is greater than 0.5, the WLAN spectrum is always very busy. Therefore, the CR system only occupies the WLAN spectrum for a very short time. On the other hand, the performance is better at low and middle offered WLAN traffic load when BUF is shortened shown in Fig. 4.6. The reason is

that the quiet periods and active periods of CR systems become smaller. Therefore, CR systems will not occupy the channel for a long time and release more opportunity to WLAN users to transmit data. As a result, CR systems with smaller basic unit frame perform better in term of the impacts of average WLAN packet delay than that with large basic unit frame.

Chapter 5 Conclusion

The limited available spectrum and the inefficiency in the spectrum usage necessitate a new communication paradigm to exploit the existing wireless spectrum opportunistically. In addition, with the demand for additional bandwidth increasing due to existing and new services, both spectrum policy and communication technologists are seeking solutions for this apparent spectrum scarcity. In this thesis, a CR-based dynamic spectrum access algorithm is proposed for CR systems downlink transmission. The white space detection in WLAN spectrum and WCDMA spectrum, frequency bands switch between two spectrums and downlink radio resource allocation are controlled by CR-based DSA algorithm. The goals of CR-based DSA algorithm are throughput maximization in a fixed bandwidth. The proposed CR-based DSA algorithm dynamically decides CR systems to be active for such periods and be quiet for another periods of time in WLAN spectrum and gives the holding time in WCDMA spectrum. In addition, it gives the timing when CR systems shift their operational spectrum to another. Finally, a simple downlink radio resource allocation is presented to maximize the throughput of CR systems.

In the simulation results and discussion, the CR-based DSA algorithm is presented for CR systems to operate in WLAN and WCDMA spectrum. From the results, we can conclude that the CR-based DSA algorithm improve the overall

resource in WCDMA spectrum.

Bibliography

[1] Federal Communications Commission, “Spectrum Policy Task Report,” ET Docket No. 02135, 2002.

[2] Federal Communications Commission, “Unlicensed Operation in the TV Broadcast Bands,” ET Docket No. 04-186, 2004.

[3] S. Haykin, Cognitive radio: Brain-empowered wireless communications, IEEE Journal on Selected Areas in Communications, vol 23, No. 2, pp 201-219, 2005.

[4] I.F. Akyildiz, W.-Y. Lee, M.C. Vuran, S. Mohanty, Next generation/dynamic spectrum access/cognitive radio wireless networks: a survey, Computer Networks (Elsevier) 50 (13) (2006) 2127–2159.

[5] M.B. Weiss,, "Secondary use of spectrum: a survey of the issues." Info the journal of policy regulation and strategy for telecommunications, 2006

[6] A. Tonmukayakul and M. B. Weiss, “Secondary use of radio spectrum: A feasibility analysis,” presented at the Telecommunications Policy Research Conference (TPRC’04), Arlington, VA, Oct. 1–3, 2004.

[7] W. D. Horne, “Adaptive spectrum access: using the full spectrum space,” 31st Annual Telecommunications Policy Research Conference (TPRC’03), October 2003.

[8] IEEE Standard Std. 802.16h Uncoordinated Coexistence Protocol (UCP)

[9] L. Nuaymi, P. Godlewski, and X. Lagrange, "Power allocation and control for the downlink in cellular CDMA networks," in Proc. 12th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC

[10] A. J. Goldsmith and S.-G. Chua, “Variable-rate variable-power MQAM for fading channels,” IEEE Trans. Commun., vol. 45, pp. 1218–1230, Oct. 1997.

[11] J. Jang and K. B. Lee, “Transmit power adaptation for multiuser OFDM system,” IEEE J. Select. Areas Commun., vol. 21, pp. 171–178, Feb. 2003.

[12] H. Yaghoobi, “Scalable OFDMA physical layer in IEEE 802.16 WirelessMAN,” Intel Techonology Journal, Volume 8, Issue 3, 2004.

[13] 3GPP TR 25.892, “Feasibility study for OFDM for UTRAN enhancement,” 3rd Generation Partnership Project, Tech. Rep., 2004-06.

Vita

Chien-Hsing Liu was born on 1982 in Taipei, Taiwan. He received the B.E.

degree in electrical and control engineering and M.E. degree in communication engineering from Nation Chiao Tung University, Hsinchu, Taiwan, in 2005 and 2007, respectively. His research interests include radio resource management and wireless communication systems.