University of Bristol - Explore Bristol Research. Link to published version (if available): /VTCF

Transcription

1 Bian, Y. Q., & Nix, A. R. (2006). Throughput and coverage analysis of a multi-element broadband fixed wireless access (BFWA) system in the presence of co-channel interference. In IEEE 64th Vehicular Technology Conference, 2006 (VTC-2006 Fall), Montreal. (pp. 1-5). Institute of Electrical and Electronics Engineers (IEEE) /VTCF Link to published version (if available): /VTCF Link to publication record in Explore Bristol Research PDF-document University of Bristol - Explore Bristol Research General rights This document is made available in accordance with publisher policies. Please cite only the published version using the reference above. Full terms of use are available: Take down policy Explore Bristol Research is a digital archive and the intention is that deposited content should not be removed. However, if you believe that this version of the work breaches copyright law please contact open-access@bristol.ac.uk and include the following information in your message: Your contact details Bibliographic details for the item, including a URL An outline of the nature of the complaint On receipt of your message the Open Access Team will immediately investigate your claim, make an initial judgement of the validity of the claim and, where appropriate, withdraw the item in question from public view.

2 Centre for Communications Research Throughput and Coverage Analysis of a Multi-Element Broadband Fixed Wireless Access (BFWA) System in the Presence of Co-Channel Interference Y.Q. Bian and A.R. Nix {y.q.bian; andy.nix}@bristol.ac.uk

3 Introduction Study builds on IEEE (16d) fixed broadband wireless access standard (fixed WiMAX) Channel model for fixed WiMAX in urban environments SUI (Stanford University Interim) Site specific model (urban ray tracing) Analyse channel and interference behaviour within a WiMAX network Work towards achieving high system throughput using the OFDM-wirelessMAN air interface by exploiting sectorised MIMO arrays Conclusions

4 Background (IEEE ) Fixed WiMAX provides a cost-effective wireless network over extensive areas to large number of users; it provides differentiated broadband service The IEEE standard defines three PHY layer air-interface techniques The 256-FFT OFDM solution is favoured by the vendor community, since OFDM improves system performance in non-line-of-sight urban environments UK has 3.5GHz licensed band WiMAX supports: Adaptive modulation and coding (AMC) Smart antennal techniques

5 SUI Channel Models (1) SUI models are widely used for fixed WiMAX deployment. A total of six models are defined for three typical terrain types SUI-1 SUI-2 SUI-3 SUI-4 SUI-5 K factor τ rms (µs ) Ant. Corre. Omni: 3.3(90%), 10.4(75%) o ant.: 14(90%), 44.2(75%) Omni: 1.6(90%), 5.1(75%) C 30 o ant.: 6.9(90%), 21.8(75%) Omni: 0.5(90%), 1.6(75%) B 30 o ant.:2.2(90%), 7.0(75%) Omni: 0.2(90%), 0.6(75%) B 30 o ant.:1.0(90%), 3.2(75%) Omni: 0.1(90%), 0.3(75%), 1.0(50%) A Terrain type C 30 o ant.:0.4(90%), 1.3(75%), 4.2(50%) SUI-6 Omni: 0.1(90%), 0.3(75%), 1.0(50%) A 30 o ant.:0.4(90%), 1.3(75%), 4.2(50%) Type A: hilly terrain with moderate-to-heavy tree densities; Type C: mostly flat terrain with light tree densities.

6 SUI Channel Models (2) Scenario for SUI channel: 1. Cell size: 7 km 2. BS ant. high: 30m 3. CPE ant. High: 6m 4. BS ant. beamwidth: CPE ant. Beamwidth: omni & Polarization Vertical 7. Cell coverage 90% with 99.9% reliability The SUI approach struggles to support link adaptation studies and is not ideal for calculating the coverage outage probability, since it makes a number of general assumptions (see above) To overcome these limitations and assumptions, we use a site specific Ray Tracing (RT) model to analyse the radio channels between each BS and their associated CPEs

7 Simulation Scenarios 3km by 1.8km area covering central Bristol (UK) 5 BS placed on tall local buildings (~30m height); each BS uses three 120 o sectorised antennas 100 CPEs are located at rooftop height (around 6m), and use omni or 30 o directional elements Consider two MIMO array configurations: 1) Uniform Circular Array (UCA), and 2) Uniform Linear Array (ULA) 5MHz channel bandwidth assumed in the 3.5GHz band

8 MIMO Channel for Sectorised Multi-BS Network Channel Processor Ray Tracing (RT) CPE-BS assignment Sector antenna RTD-MIMO model H Antenna pattern Array geometry Utilises urban geographic data to produce spatial/temporal channel using isotropic element patterns. ETSI specified beam patterns are then spatially convolved to model the impact of directional elements CPE to BS-Sector assignment and orientation based on strongest path Dominate ray CPE BS Dominate ray CPE

9 Ray Tracing Deterministic (RTD) - MIMO Channel (1) SISO RT tool Antenna pattern superimposing Spatial correlations are determined by the spatial directions of the multipaths (AoD/AoA) and the array geometry Frequency correlations are controlled by channel time delay spread. Amp. & phase AoD & AoA delay RTD-MIMO Q Phase shifter H n,m generator NR NT Q Time binning DFT N N L N R NT N f R T H w array geometry bandwidth

10 RTD - MIMO Channel (2) The plot below shows the time varying Tapped Delay Line weights, and the resulting frequency selective channel responses for the MIMO links Each MIMO link suffers individual frequency selective fading

11 Comparison between RT and SUI model The RT results demonstrate that: - 73% of users have a very strong dominant ray (a K-factor above 15dB). Simulated area fits the SUI-3 assumptions (terrain type B) -30 o directional antenna increases the K-factor - RMS Angle Spread is much lower in Elevation (compared to Azimuth) Probility of K-factor>abscissa SUI-3 RT: omni CPE SUI-1: omni CPE SUI-2: omni CPE SUI-3: omni CPE SUI-4: omni CPE SUI-5: omni CPE SUI-6: omni CPE RT: 30 deg. CPE SUI-1: 30 deg. CPE SUI-2: 30 deg. CPE SUI-3: 30 deg. CPE SUI-4: 30 deg. CPE SUI-5: 30 deg. CPE SUI-6: 30 deg. CPE No. of CPE Azimuth Elevation BS: 120-sector antenna RMS angle spread (degree)) 43% CPE only has single traced ray K-factor (db)

12 Array Element Geometry Higher spatial correlation seen between vertically displaced array elements. Consequently, it is more effective to space MIMO elements in the horizontal plane The figure below shows the structure of the MIMO sectorised array element-3 (φ =90 0 ) sector-2 element-4 (φ = ) ρ AoD Sector-3 sector-1 element-2 (φ = 0 0 ) N W S E element-1 (φ = )

13 Interference in multi BS environment (1) SINR C = log 2 det I N + HH R NT H Co-Channel Interference (CCI) has a harmful effect on capacity (depending on the frequency reuse plan) Key challenges include improving throughput, spectrum efficiency and coverage BS-6 BS-7 f 7 = f 1 BS-1 f 6 BS-5 f 1 BS-4 f 5 = f 1 f 4 BS-2 f 2 = f 1 BS-3 f 3 = f 1

14 C/I Distributions Systems with directional antenna are likely to see high coverage through the use of higher link-speeds Performance of Interference Cancellation (IC) depends on the specific channel characteristics 14 Omni elements o directional elements Occurences 8 6 Occurences Carrier to interference ratio (db) Carrier to interference ratio (db)

15 Interference in multi BS environment (2) MIMO spatial correlation further reduces capacity 40 2x2 MIMO capacity for case of CCI 35 no CCI 30 C/I = 30dB Capacity (bit/sec/hz) H uncorrelated 2 1 = j j 1 C/I = 3dB C/I = 20dB C/I = 10dB no CCI C/I = 10dB C/I = 30dB C/I = 20dB 10 H 1 1 = 1 C/I = 3dB 1 correlated SNR (db)

16 Array Geometry Impact on MIMO Performance RCN values quantify MIMO spatial correlation and reflect channel capacity ULA geometry works well when the channel is characterized by low correlation When MIMO channels are highly correlated, UCA structures outperform their ULA counterparts capacity (bit/s/hz) x4 ULA omni RCN 4x4 MIMO channel RCN for ULA omni antennas 4x4 MIMO channel RCN for UCA omni antennas RCN: Reciprocal Condition Number ULA: Uniform Linear Array UCA: Uniform Circular Array AP AP

17 Results Analysis In the presence of CCI, there is a trade-off between interference suppression (using narrow beamwidths) and diversity gain (using wider beamwidths). We recommend: - Wide MIMO element spacings (e.g. 5λ for the BS and 0.5λ for the CPE ); - Deploying MIMO arrays in the horizontal plane. UCA (4x4 MIMO) works well for open areas. Compared to SISO, the sectorised 2x2 MIMO system improved coverage of the highest link speed by 12%. Hence system throughputs are improved by 1.5Mbps. 44% SISO system throughputs: 11.4Mbps 56% 2x2 MIMO system throughputs: 12.9Mbps 1/2 BPSK 1/2 QPSK 3/4 QPSK 1/2 16-QAM 3/4 16-QAM 2/3 64-QAM 3/4 64-QAM

18 Conclusions Radio channel plays an important role in system evaluation. Our RTD-MIMO can simulate propagation across transition regions, and thus presents a unique insight into system coverage and throughput. We have demonstrated WiMAX performance in a practical urban environment, and shown degradation due to path loss, spatial correlation and inference. Our results show that: Sectorised antenna system consistently offers higher C/I than omni antenna equivalents. However, they still cannot achieve full coverage and high throughputs on their own Sectorised MIMO arrays offer significant benefits (i.e. the 1.5Mbps improvement seen in this study), even with many LoS channels