Joint Cooperation Design Simulation Results AnalysisAnalysis

Simulation results of different CoMP schemes in both the intra and inter-site are provided in this section. The average throughput of SU-MIMO in the entire cell and at the cell edge is 2.39 (bps/Hz/cell) and 0.0791 (bps/Hz/user), respectively. We regard the performance of SU-MIMO as a baseline and compare it with different CoMP schemes in Figs. 5.3 and 5.4. When adopting JP in the intra-site technique with RRH location in 0.6R (R is the cell radius), spectrum efficiency improve 23%.

The proposed scheme intra-JP and inter-CS/CB scheme can further improve spectrum efficiency by 41.7% at the cell edge and 15% in the entire cell. Thus, the proposed joint design of CoMP transmission scheme can effectively mitigate the ICI on average.

According to our previous work, we provide the performance of MU-MIMO in cell edge and cell average, which are 0.086 bps/Hz/user and 2.62 bps/Hz/cell. In Figs.

5.5 and 5.6, the SU/MU-MIMO switching techniques are adopted. Since the serving cell can connect multiple users in the same resource block by exploiting the degrees of freedom in the spatial domain, MU-MIMO can achieve higher throughput than only adopting SU-MIMO. As shown in Figs. 5.5 and 5.6, the proposed scheme can achieve additional gain of 54.65% and 19.32% in cell edge and cell average compared to the inter-JP CoMP. As compared with SU-MIMO, the proposed joint design, we can have relative gain of 100.63% and 61.24%, which is very significant.

We consider the effect of cell architectures on SU-MIMO and CoMP schemes.

served by macro-cell and RRHs. Compared with pentagonal, the proposed scheme has performance gain of 8% and 3.32% in cell edge and cell average, respectively.

In Figs. 5.9 and 5.10, phase compensation techniques can further enhance the performance of three different CoMP schemes, especially for inter-JP plus intra-JP transmission scheme. Compared to SU-MIMO, the proposed CoMP joint design technique improves performance of 71.6% and 49% in the cell edge and cell average, respectively. We listed the performance of three different CoMP schemes in Table 5.11. Based on [3], the feedback delay of 20 msec will degrade the system performance by 5% and 10% in cell average and cell edge, respectively. As shown in Table 5.11, our results can match the results provided by the 3GPP standard.

We use different transmission schemes in this chapter to show that the joint cooperation design can outperform other existing methods. In this section, we com-pare the performance of intra-site JP combined with inter-site CS/CB and inter and intra-site JP. The former transmission scheme is more efficient than the latter. We suggest that when operating CoMP techniques in both sites, the best combination method is the intra-site JP and the inter-site CS/CB. 3GPP recommends three types of cell architectures. We investigate the effects of differences between architectures upon our approach and other CoMP techniques.

34

Table 5.1: Simulation parameters for MIMO systems.

Parameter Value

Duplex Method FDD

DL Transmission Scheme OFDMA

Subcarrier Number 600

Downlink Transmit 2 and 4

Antenna Number

Downlink Receive 2

Antenna Number

ISD 500 meters

Macro-cell Number 19

Number of Users per Sector 10

User Speed 3 km/hr

Network Synchronization synchronization Downlink Scheduler proportional fair scheduling

in time and frequency

Downlink ARQ a maximum of four transmission times

Downlink Receiver Type MMSE

BS Transmit Power 46 dBm

Noise Power Density −174 dBm/Hz Antenna Configuration 0.5 wavelength separation

at Receiver

Antenna Configuration 10 wavelength separation at Transmitter

Minimum Distance between 35 meters user and macro-cell

Channel Model SCM urban macro

Antenna Pattern AH(ϕ) =− min[ Shadowing Model lognormal with zero mean and

8 dB standard deviation

36

Figure 5.2: Minimum distance between devices.

Figure 5.3: Comparisons of different CoMP schemes at the cell edge.

38

Figure 5.4: Comparisons of cell average performance for different CoMP schemes.

Figure 5.5: Comparisons of different CoMP schemes at the cell edge versus SU/MU-MIMO switching.

40

Figure 5.6: Comparisons of different CoMP schemes in the cell average versus SU/MU-MIMO switching.

Figure 5.7: CoMP schemes performance with different cell architectures at the cell edge.

42

Figure 5.8: CoMP schemes performance differences for different cell architectures in the entire cell.

Figure 5.9: Effects of phase compensation for different CoMP schemes at the cell edge.

44

Figure 5.10: Effects of CoMP schemes performance on the cell average.

46

CHAPTER 6

Conclusions

6.1 Summary

We operate coherent JP in heterogeneous networks based on 3GPP LTE standard [3] and obtain a better performance than non-coherent JP. However, it needs more information about channel state and leads to larger overhead. Now we consider delay effect with over two frames of delay time, which degrades performance significant.

Based on the ICI mitigation beamforming design, we present a joint cooperation design of CoMP techniques in both the intra- and inter-site, and it is a feasible method to further improve the system performance. When we try to extend the cooperative range from the intra-site to inter-site, we find that the CS/CB transmission techniques in the inter-site is more effective and efficient than JP. Since this approach provides more relative gain and needs less backhaul requirement, we suggest that the joint cooperation design can be JP in intra-site and CS/CB in inter-site. The narrow-beam tri-sector architecture is the most suitable cell architecture for the inter-CS/CB plus intra-JP CoMP scheme. Our proposed scheme needs less feedback information and can further enhance the spectrum efficiency.