4.4 DCC Design
4.5.2 Performance Measurements and Discussions
We compare the performance of four schemes, including fixed pilot with SSDT (FIX-SSDT), LPPA (FIX-LPPA) schemes, and dynamic cell configuration with SSDT (DCC-SSDT), LPPA (DCC-LPPA) schemes. For the fixed pilot scheme, the default pilot power, PbI, is fixed and set as 2.5 watt for each cell (12.5% of maximum transmission power).
1 1.2 1.4 1.6 1.8 2 2.2 2.4
Figure 4.8: For fixed pilot power design, capacity results by applying SSDT and LPPA schemes in terms of different referenced service coverage under uniform (ρ = 1) and non-uniform (ρ = 4) cell load cases.
The maximum link power, epb, and the threshold of the call admission control, HS, are fixed and calculated from (4.38) and (4.40), respectively. The DCC scheme is the proposed reinforcement-learning-based dynamic cell configuration scheme. For the DCC scheme, PbI, e
pb, and bHS are adjusted dynamically as described in section 4.4. Assume the arrival rate is 1.6, and the traffic load ratio, ρ, is changing from 1 to 5. For the design parameters of the DCC scheme, maximum and minimum fractions of pilot power are fmin = 0.05 and fmax = 0.2, respectively; decision period N is 10 frames (100 msec); total number of measurement samples M is 10 frames; total simulation time is 106 frames (105 learning times).
The comparison between FIX-LPPA and FIX-SSDT in terms of capacity and coverage is shown in Fig. 4.8. We see that the FIX-LPPA scheme achieves higher total throughput than the FIX-SSDT scheme for both uniform and non-uniform cell load cases. This is because FIX-LPPA successfully releases congested cell’s load through power balance strategy so that it can support higher throughput than the FIX-SSDT scheme around 20% in the non-uniform cell load case. On the other hand, the FIX-SSDT scheme lacks of flexibility to adapt to non-uniform cell load situations so that the throughput is decreased compared to the non-uniform
1 2 3 4 5
Figure 4.9: The average pilot power of hotspot, 1st-tier, and 2nd-tier cells for the fixed pilot (FIX) with SSDT (FIX-SSDT) and LPPA (FIX-LPPA), and pilot power allocation for dynamic cell configuration (DCC) with SSDT (DCC-SSDT) and LPPA (DCC-LPPA) schemes.
cell load case. It is noteworthy that the ripple occurs when the simulation result of reference rate equals 6. This is because the simulation time is not enough for accommodating high rate services which are interfere by many algorithms of radio resource management, i.e. call and handoff admission control, handoff algorithm, etc.
Figure 4.9(a) and (b) show the average pilot power distribution of the hotspot, 1st-tier, and 2nd-tier cells using DCC-LPPA and DCC-SSDT schemes, respectively. We can see that the DCC scheme adjusts the pilot power in each cell according to various system situations. When the traffic load ratio is getting heavier, the pilot power of the hotspot cell is reduced aggressively so as to balance traffic load to adjacent cells, but the coverage is shrunk accordingly. In this way, the base station of hotspot can save more transmission power for traffic channels to serve new call arrivals. Besides, adjustments of the pilot power can make the existing mobile stations near the cell boundary enter soft handoff mode so as to balance traffic load. Furthermore, for hotspot cell, the slope of the pilot power level of different traffic load ratio for DCC-SSDT is sharper than that for DCC-LPPA. This is
1 2 3 4 5
Figure 4.10: The average blocking probability of (a) real-time and (b) non-real-time services for FIX-SSDT, FIX-LPPA, DCC-SSDT, and DCC-LPPA schemes, respectively.
because both DCC and LPPA strategies are helpful for power balance; the pilot power of DCC-LPPA does do not have to be adjusted as aggressive as DCC-SSDT.
Figure 4.10(a) and (b) show the new call blocking probabilities of real-time and non-real-time services, respectively. We can see that the DCC scheme improves the blocking probabilities of the FIX scheme for both real-time and non-real-time services. Because the base station has limited transmission power resources, it is important to achieve power balance between cells in downlink transmission. In order to achieve power balance between cells, our proposed DCC scheme adjusts pilot power and coordinates other radio resource mechanisms dynamically. This is the reason why the DCC scheme can save more power resource to accommodate new call requests. It is noteworthy that LPPA and DCC-SSDPT schemes without changing the other radio resource management criteria are also presented for comparison. We can see that the DCC scheme without coordinating other radio resource management criteria has worse new call blocking probability performance than the FIX scheme. This is because when new or handoff calls issue requests to cells with light (heavy) traffic load, the tight (loose) criteria of call admission control may result in
1 2 3 4 5 10−4
10−3 10−2 10−1
Traffic load ratio ( ρ )
Average handoff forced termination probability
FIX−SSDT FIX−LPPA DCC−SSDT DCC−LPPA
DCC−SSDT & FIX−RRM DCC−LPPA & FIX−RRM
Figure 4.11: The average handoff forced termination probability for FIX-SSDT, FIX-LPPA, DCC-SSDT, and DCC-LPPA schemes.
new call blocks.
The similar impaired results occur when a mobile station fails to add new base stations into its active set and suffers worse channel quality, as shown in Fig. 4.11. This is so-called handoff forced termination. On the other hand, compared to the FIX scheme, our proposed DCC scheme can improve handoff forced termination probabilities greatly. This is because existing mobile stations near the cell boundaries often suffer bad transmission quality, and they may be dropped when power resource is not enough for admitting handoff requests.
Figure 4.12 shows the total throughput of the system. In the FIX case, FIX-LPPA outperforms FIX-SSDT. When the traffic load ratio is getting higher, the throughput of FIX-SSDT degrades sharply because of the inefficient handoff power allocation strategy. As for FIX-LPPA, it can even keep the average throughput when traffic load ratio is less than 4. Furthermore, through the cooperation of DCC and LPPA schemes for power balance between cells, DCC-LPPA can further enhance FIX-LPPA. Compared to the FIX scheme, the DCC scheme can improve the average throughput when the traffic load ratio is getting larger for both DCC-SSDT and DCC-LPPA schemes. This is because the DCC scheme can
1 2 3 4 5 240
250 260 270 280 290 300 310 320 330 340
Traffic load ratio ( ρ )
Total throughput (frames)
FIX−SSDT FIX−LPPA DCC−SSDT DCC−LPPA
Figure 4.12: The average total throughput of systems for FIX-SSDT, FIX-LPPA, DCC-SSDT, and DCC-LPPA schemes.
1 2 3 4 5
10−4 10−3 10−2 10−1
Traffic load ratio ( ρ )
Average frame error rate
FIX−SSDT FIX−LPPA DCC−SSDT DCC−LPPA
Figure 4.13: The average frame error probability for FIX-SSDT, FIX-LPPA, DCC-SSDT, and DCC-LPPA schemes.
1 2 3 4 5 1.2
1.25 1.3 1.35
Traffic load ratio ( ρ )
Average active set size
FIX−SSDT FIX−LPPA DCC−SSDT DCC−LPPA
Figure 4.14: The average size of the active set for FIX-SSDT, FIX-LPPA, DCC-SSDT, and DCC-LPPA schemes.
dynamically balance traffic load between cells through pilot power adjustments based on system situations as well as call admission control criterion and the maximum link power constraint.
Furthermore, Fig. 4.13 presents the results of frame error probabilities. We observe that frame error probability can be improved by DCC schemes, and the requirement of frame error probability 0.01 can be satisfied through an effective feature abstraction design.
In order to balance traffic loads between cells, the DCC scheme can reduce or increase pilot power aggressively. The power balance can be achieved by forcing mobile stations near the cell boundary into handoff mode. Therefore, the average size of the active set and handoff rates can be increased, which is shown in Fig. 4.14. It is found that the DCC scheme results in more soft handoff events slightly.
Furthermore, Table 4.2 shows the results of coverage failure probabilities. A coverage failure occurs when a new call fails to detect good enough signal from a base station for requesting a traffic channel. From equation (4.15), we can see that there are two situations that could cause a coverage failure. One is when the mobile station suffers high interferences (the increment of the denominator in (4.15)), and another is when the mobile station has
weak signal strength (the reduction of the nominator in (4.15)). The SSDT and DCC-LPPA schemes cause slightly higher coverage failure probabilities than the FIX schemes. This is because even DCC devotes to balance traffic load through the pilot power adjustments so as to reduce interference of the hotspot cell, mobile stations near the cell boundary may suffer bad signal strength from any base station in the active set. This is the possible disadvantage for the DCC schemes in the distributed manner.
Table 4.2: Average coverage failure probability Scheme/Traffic load ratio 1.0 2.0 3.0 4.0 5.0
FIX-SSDT 0.0 0.0 0.0 0.0 0.0
FIX-LPPA 0.0 0.0 0.0 0.0 0.0
DCC-SSDT 0.0 0.0 4.2e-05 1.0e-04 4.6e-04 DCC-LPPA 0.0 0.0 0.0 3.2e-05 9.2e-05
4.6 Concluding Remarks
In this chapter, we have studied dynamic cell configuration problem in next generation CDMA networks. The large number of states and the difficulty to estimate the state tran-sition probabilities in realistic networks motivate us to choose a model-free reinforcement-learning solution to solve the problem. The proposed scheme can dynamically configure cell coverage and capacity based on the varying situation of the systems. Simulation results have been shown the effectiveness of the proposed schemes. It was found that pilot and soft handoff power allocations, maximum power constraint design, and the admission con-trol criterion are highly coupled and should be considered jointly. The system throughput can be increased significantly in the proposed dynamic cell configuration scheme compared to the conventional fixed scheme. Furthermore, both dynamic cell configuration and link proportional power allocation schemes have the advantage of power balance for soft handoffs so that the system capacity of the DCC-LPPA scheme outperforms conventional FIX-SSDT scheme significantly.
The proposed dynamic cell configuration scheme is the initiative framework for the next-generation CDMA networks. The dynamic cell configuration concept is applicable to future
cellular systems with whatever embedded multiple access techniques. Advanced study is in progress to apply the DCC scheme to the possible mechanisms of the next-generation systems.
Chapter 5
Conclusions and Future Works
In this dissertation, we specialize in the downlink soft handoff mechanisms and cell re-configuration planning in terms of power balance characteristics to tackle tradeoffs between service coverage and system capacity in CDMA mixed-size cellular systems. We first con-sider a CDMA mixed-size cellular system with mixed-size cells in Chapter 2 and 3, in which the cell configuration is determined by fixed pilot power allocation. Radio resource man-agement of soft handoff for the narrowband CDMA supporting voice only system and the multirate WCDMA system are considered in Chapter 2 and 3, in which we propose a novel soft handoff power allocation scheme and joint power and rate allocation schemes, respec-tively. The snapshot simulations are adopted to evaluate the qualitative characteristics for the CDMA heterogenous cellular system. Next, in Chapter 4, consider the cellular system with time-varying non-uniform cell loads distribution, we further design a reinforcement-learning-based dynamic cell configuration scheme with radio resource management to be aware of system situation and to adjust pilot power, maximum link power as well as call admission criterion dynamically. A practical wireless mobile cellular environment has been set up to simulate the dynamic cellular system with random mobility and versatile services activity. The conclusions of this dissertation are highlight as follows.
In Chapter 2, we propose a novel link proportional power allocation (LPPA) scheme and compare it with many existing soft handoff power allocation schemes, including EPA, SSDT, and QBPA schemes. In the simulations, we show that LPPA can effectively relax the power exhausting problem. Specifically, by taking into account effects of different cell sizes, LPPA can prevent a microcell’s base station from wasting too much transmission power in serving
handoff users. Consequently, the LPPA scheme can deliver higher system capacity and service coverage than other soft handoff power allocation schemes in both the homogeneous and mixed-size cellular systems even with measurement errors. All in all, we find that it is important to design a handoff mechanism from perspectives of power efficiency and link reliability in the CDMA mixed-size cellular system with mixed-size cells.
In Chapter 3, a joint power and rate assignment (JPRA) algorithm has been proposed to deal with multirate soft handoffs in WCDMA mixed-size cellular systems, containing the LPPA scheme and the evolutionary computing rate assignment (ECRA) method. Compared to SSDT and LPPA schemes with best-effort based rate assignments, simulation results show that JPRA accomplishes superior power balance among cells so that the system performance can be improved significantly, including better cell coverage, higher system throughput, and great user satisfaction for voice and data users. Furthermore, JPRA is less sensitive to the measurement errors during active set selection than SSDT with best-effort rate allocation, so JPRA also owns better link reliability. It is noteworthy that the aforementioned advantages of JPRA are more conspicuous in WCDMA mixed-size cellular systems with smaller cell radius ratio between microcell and macrocells.
In Chapter 4, we have studied dynamic cell configuration problem in next generation CDMA networks. The large number of states and the difficulty to estimate the state tran-sition probabilities in realistic networks motivate us to choose a model-free reinforcement-learning solution to solve the problem. The proposed scheme can dynamically configure service coverage and system capacity based on the varying situation of the systems. It is found that pilot and soft handoff power allocations, maximum link power constraint as well as the admission control criterion are highly coupled and should be considered jointly.
Simulation results have been shown the effectiveness of the proposed schemes. The system throughput can be increased significantly in the proposed dynamic cell configuration scheme with radio resource management compared to the conventional fixed scheme. Furthermore, the system capacity of the dynamic cell configuration with LPPA (DCC-LPPA) scheme out-performs conventional fixed cell configuration with SSDT (FIX-SSDT) scheme significantly.
It means that the DCC-LPPA scheme owns the excellent advantage of power balance. There-fore, for future multimedia and personal communications, cell area will be further reduced, and traffic variation and unevenness will be expanded. Dynamic cell configuration and soft handoff mechanism to balance cell loads are crucial to enhance the system efficiency for all types of cellular systems.
The proposed dynamic cell configuration scheme with radio resource manage is the ini-tiative framework for the next-generation CDMA networks. The dynamic cell configuration concept is applicable to future cellular systems with whatever embedded multiple access techniques. Advanced study is to apply it to the possible mechanisms of the next-generation systems. Furthermore, in order to provide more flexible plan of cell reconfiguration, future works can take into account uplink and downlink capacity-limited scenarios together. Since the volumes of multimedia traffic loads are greatly asymmetrical, the dynamic cell configura-tion scheme with radio resource management affects both links of their service coverage and system capacity differently. The possible challenge of applying reinforcement-learning-based techniques distributively to the bidirectional dynamic cell configuration would be that there may induce some problems of the interaction between uplink and downlink transmissions for different cells. The optimal decisions of the pilot power level for the bidirectional dynamic cell configuration would face convergent problem. Some advanced multi-agent reinforcement-learning techniques would be useful to tackle the interaction problems.
Bibliography
[1] J. Laiho, A. wacker, and T. Novosad, Eds, Radio Network Planning and Optimization for UMTS, Wiley & Sons, pp. 280-290, 2002.
[2] K. S. Gilhousen, et al., “On the capacity of a cellular CDMA system,” IEEE Trans.
Veh. Technol., vol. 40, no. 2, pp. 303-312, May 1991.
[3] W. C. Y. Lee, “Overview of cellular CDMA,” IEEE Trans. Veh. Technol., vol. 40, no.
2, pp. 291-302, May 1991.
[4] W. W. Lu, Broadband wireless mobile: 3G and beyond, John Wiley & Sons Ltd, pp.
307-315, 2002.
[5] H. G. Jeon, S. M. Shin, T. Hwang, and C. E. Kang, “Reverse link capacity analysis of a CDMA cellular system with mixed cell sizes,” IEEE Trans. Veh. Technol., vol. 49, no.
6, pp. 2158-2163, Nov. 2000.
[6] S. Kishore, L. J. Greenstein, H. V. Poor, and S. C. Schwartz, “Uplink user capacity in a CDMA macrocell with a hotspot microcell: exact and approximate analyses,” IEEE Trans. Wireless Commun., vol.2, no.2, pp. 364-374, Mar. 2003.
[7] S. Kishore, L. J. Greenstein, H. V. Poor, and S. C. Schwartz, “Downlink user capacity in a CDMA macrocell with a hotspot microcell,” in Proc. IEEE GLOBECOM’03, San Francisco, CA, pp. 1573-1577, Dec. 2003.
[8] C. Y. Liao, L. C. Wang, C. J. Chang, “Power allocation mechanisms for downlink hand-off in WCDMA systems with heterogeneous cell structures,” ACM/Kluwer WINET, Jun. 2005.
[9] A. J. Viterbi, CDMA: priciples of spread spectrum communication, Addison-Wesley, pp.
218-224, June 1995.
[10] H. Holma, and A. Toskala, “WCDMA for UMTS: radio access for third generation mobile communications,” John Wiley and Sons, ltd., pp. 208-210, 2000.
[11] O. Salonaho, and R. Padovani, “Flexible power allocation for physical control channel in wideband CDMA,” IEEE VTC’99 Spring, Houston, TX USA, pp. 1455-1458, May 1999.
[12] L. Dai, S. D. Zhou, and Y. Yao, ”Effect of macrodiversity on CDMA forward-link capacity,” IEEE VTC’01 Fall, Atlantic, NJ USA, pp. 2452-2456, 2001.
[13] D. Kim, “A simple algorithm for adjusting cell-site transmitter power in CDMA cellular systems,” IEEE Trans. on Veh. Technol., vol. 48, no. 4, pp.1092-1098, July 1999.
[14] H. Furukawa, K. Hamabe, and A. Ushirokawa, “SSDT − Site Selection Diversity Trans-mission Power Control for CDMA Forward Link,” IEEE J. Selected Areas in Commun., vol. 18, no. 8, pp. 1546-1554, Aug. 2000.
[15] N. Takano, and K. Hamabe, “Enhancement of site selection diversity transmit power control in CDMA cellular systems,” IEEE VTC’01 Fall, Atlantic, NJ USA, pp. 635 -639, Oct. 2001.
[16] F. Blaise, L. Elicegui, F. Goeusse, G. Vivier, “Power control algorithms for soft handoff users in UMTS,” IEEE VTC’02 Fall, Vancouver, BC Canada, pp. 1110-1114, 2002.
[17] D. Staehle, K. Leibnitz, and K. Heck, “Effects of soft handoff on the UMTS downlink perpormance,” IEEE VTC’02 Fall, Vancouver, BC Canada, pp. 960-964, 2002.
[18] S. L. Kim, Z. Rosberg, and J. Zander, “Combined power control and transmission rate selection in cellular networks,” in Proc. IEEE VTC’99 Fall, Amsterdan, Netherlands, pp. 1653-1657, Sept. 1999.
[19] C. W. Sung, W. S. Wong, ”Power control and rate management for wireless multimedia CDMA systems,” IEEE Trans. Commun. vol. 49, no. 7, pp. 1215-1226, July 2001.
[20] D. I. Kim, E. Hossain, and V. K. Bhargava, “Downlink joint rate and power allocation in cellular multirate WCDMA systems,” IEEE Trans. Wireless Commun., vol. 2, no. 1, pp. 69-80, Jan. 2003.
[21] S. Kahn, M. K. Gurcan, and O. O. Oyefuga, “Downlink throughput optimization for wideband CDMA systems,” IEEE Commun. Lett., vol. 7, no. 5, pp. 251-253, May 2003.
[22] D. Kim, ”Rate-regulated power control for supporting flexible transmission in future CDMA mobile networks,” IEEE J. Selected Areas in Commun., vol. 17, no. 5, pp.
968-977, May 1999.
[23] Y. W. Kim, D. K. Kim, J. H. Kim, S. M. Shin, and D. K. Sung ”Radio resource management in multiple-chip-rate DS/CDMA systems supporting multiclass services,”
IEEE Trans. Veh. Technol., vol. 50, no. 3, pp. 723-736, May 2001.
[24] D. K. Kim, and D. K. Sung, ”Handoff management in CDMA systems with a mixture of low rate and high rate traffics,” in Proc. IEEE VTC’98 Spring, Ottawa, ON, pp.
1346-1350, May 1998.
[25] S. A. Grandhi, J. Zander, and R. D. Yates, “Constrained power control,” Wireless Personal Commun., vol. 1, no. 4, pp. 257-270, 1995.
[26] M. Andersin, Z. Rosberg, and J. Zander, “Gradual removals in cellular PCS with con-strained power control and noise,” ACM/Baltzer Wireless Networks J., vol. 2, no. 1, pp. 27-43, 1996.
[27] F. Berggren, R. Jantti, and S. L. Kim, “A generalized algorithm for constrained power control with capability of temporary removal,” IEEE Trans. Veh. Technol., vol. 50, no.
6, pp. 1604-1612, Nov. 2001.
[28] S. L. Kim, “Optimization approach to prioritized transmiter removal in a multiservice cellular PCS,” in Proc. IEEE PIMRC’98, Boston, MA, pp. 1565-1569, Sept. 1998.
[29] S. Sharma, A. R. Nix, and S. Olafsson, “Situation-aware wireless networks,” IEEE Commun. Mag., pp. 44-50, July 2003.
[29] S. Sharma, A. R. Nix, and S. Olafsson, “Situation-aware wireless networks,” IEEE Commun. Mag., pp. 44-50, July 2003.