Transforming MIMO Test MIMO channel modeling and emulation test challenges Presented by: Kevin Bertlin PXB Product Engineer Page 1 Outline Wireless Technologies Review Multipath Fading and Antenna Diversity STC and MIMO Background Channel Models and Spatial Correlation MIMO Channel Emulation Page 2 1
Major Shifts in Wireless Technology More capacity on the same spectrum = More revenue for ISPs 1980 s 1990 s 2000 s Frequency Time Space AMPS GSM W-CDMA Analog Modulation FM Digital Modulation QPSK, OFDMA Multiple Antennas WiMAX, 802.11n, LTE Page 3 Multiple Antenna Configurations SISO MISO Tx Rx Tx0 Rx Tx1 SIMO MIMO Tx Rx0 Tx0 Rx0 Rx1 Tx1 Rx1 Page 4 2
Wireless Standards and MIMO Multi-Antenna Technologies Going Mainstream WiMAX Tx0 h00 Rx0 LTE 802.11n WLAN Tx1 h01 h10 h11 Rx1 Agilent is leading the MIMO evolution in wireless comms Page 5 Comprehensive MIMO Test Solutions Leading the Evolution in Wireless Comms Leader in MIMO Signal Creation Best ACPR performance for amplifier test Synchronized & phase coherent signals for MIMO receiver test N5182A MXG Tx0 Superior Signal Analysis Capabilities Single and multi-channel measurements Infiniium Accelerate Protocol Development & Conformance Real-time LTE and WiMAX base station emulation for mobile development Tx1 E6620A N5106A PXB MIMO Receiver Tester N9020A MXA 89600 VSA software E6651A Page 6 3
SISO and Multipath Fading All wireless transmissions experience signal fading Base Station SISO Subscriber Received signal is uniquely defined between the pair of antennas and the environment λ 2 Page 7 Multiple Antennas Improve Signal Quality Antenna Base Station #1 #2 Antenna Diversity Improves Received Average SNR Subscriber Page 8 4
Spatial Diversity Techniques Switched Diversity Beamforming Maximal Ratio Combining Transmit Diversity Page 9 Improving Signal Quality: Space Time Coding #1 #2 Decoder Receiver measures Symbol recovery requires two symbol times Requires knowledge of channel (h 0 & h 1 ) Page 10 5
Multiple Antennas Improve Capacity Tx0 Transmit Antennas Receive Antennas Rx1 Multipath Increases Number of Unique Spatial Channels Rx0 Tx1 Base Station MIMO Subscriber SISO MIMO S C : B log 2 1+ N S C : k B log2 1+ N k equals # of antenna pairs Page 11 Improving Data Capacity: MIMO Decoder Receiver measures Requires knowledge of channel (h 00, h 01 h 10, & h 11 ) Two symbols recovered at same time Page 12 6
Measuring Channel Matrix Coefficients [H] WiMAX Wave 2 Example Pilot subcarriers don t overlap Page 13 Measuring Channel Matrix Coefficients [H] WiMAX Wave 2 Example Pilot subcarriers don t overlap for the n th symbol Tx0 Measure pilot subcarriers Tx1 [H] Knowing s 0 and s 1, calculate [H] Page 14 7
Correlation in Channel Matrix [H] MIMO only works when the channel is rich in multipath (low channel correlation) In practice h h h h Worst case: all paths fully correlated r = r = h ( s + s ) 0 1 0 1 One equation, two unknowns Mathematically 1 [ S] = [ H ] [ R] If h coefficients are correlated, then [H] is ill-conditioned and difficult to invert Page 15 Emulating MIMO Channels Traditional Models Spatial Models The channel emulator requires accurate fading models Delay Spread Doppler Spread Angular Spread Antenna Pattern Antenna Spacing Antenna Polarization Agilent N5106A PXB MIMO Receiver Tester Page 16 8
Delay Spread Channel Model Model the channel as a tapped delay line Page 17 Doppler Spread Channel Model f d = f c v c Classical 3dB Classical 6dB Doppler Spectrum Flat Rounded Page 18 9
Improved Channel Models: Spatial Correlation Spatial Correlation exists due to Angular Spread Antenna Spacing Antenna Pattern Antenna Polarization θ angle of departure θ angle of arrival d a Page 19 Why Model Angular Spread? In general, the angular spread at the base station is less than the spread at the subscriber Page 20 10
Power Azimuth Spectrum Measured Model Cluster #2 1.2 Power Cluster #1 +π Power (linear) 1 0.8 0.6 0.4 0.2 Gaussian Uniform Laplacian π Angle (radians) +180 0 0 -pi Angle ( radian) pi Model the power distribution as a function of angular spread Page 21 Antenna Spacing and Antenna Pattern Antenna Spacing Antenna Pattern Dipole Antenna Azimuth Polar Plot Page 22 11
Spatial Correlation Low Angular Spread Large Angular Spread Antenna spacing Page 23 Antenna Polarization High Correlation Configurations Horizontal Polarization BS spacing 4λ o o polarization: 0,0 Vertical Polarization MS spacing 0.5λ o o polarization: 0,0 Low Correlation Configurations BS spacing 4λ o o polarization: -45, + 45 MS spacing 0.5λ o o polarization: 0,90 Page 24 12
Capacity Loss from Spatial Correlations Capacity (bps/hz) Uncorrelated Correlated Loss from Correlations SNR (db) α = spatial correlation coefficient at transmitter β = spatial correlation coefficient at receiver Page 25 MIMO Channel Emulation Configurations Fading Baseband Fading Apply to h xx Delay Spread Doppler Spread PAS Antenna Pattern Antenna Spacing Antenna Polarization Software Fading Page 26 13
PXB MIMO Receiver Tester Breakthrough Technology Takes MIMO Forward Transforming MIMO Test Best way to minimize WiMAX and LTE design uncertainties and rework Save hours on equipment and lab setup Performance & scalability to meet your future testing needs Page 27 Best Way to Minimize Design Uncertainties & Rework Problem Solved: R&D engineers need to test real-world MIMO and BB scenarios in the lab beyond the standard s requirements and validate performance against simulations more quickly and accurately Best-in-class Baseband Performance 120 MHz modulation bandwidth 512 Msa playback memory per BBG Up to four baseband generators Diversity test Baseband summing Standards Compliant Signal Creation Signal Studio in the PXB LTE (TDD and FDD), WiMAX, W-CDMA, GSM/EDGE Coming soon: 802.11n WLAN, digital video Advanced Channel Emulation 2x2, 2x4, and 4x2 MIMO Up to 120 MHz fading bandwidth Up to 24 paths per fader Single-user and multi-user MIMO Comprehensive Channel Correlation Settings Standards-based channel models Antenna setup Spacing Radiation pattern Custom correlation matrix Page 28 14
Save Hours on Equipment and Lab Setup Problem Solved: Time consuming re-cabling to for different HW configurations Reconfigure in seconds DSP blocks Supports up to 12 DSP blocks Each DSP block is FW configurable Baseband Generator Real-time fader Reconfigurable Architecture with Internal Signal Routing I/O ports Supports up to eight I/O ports Each I/O port Analog I/Q outputs Connectivity to external instrument N5102A for digital output MXG/ESG signal generator MXA signal analyzer Page 29 Save Hours on Equipment and Lab Setup Problem Solved: When using faders calibrations can take hours or days Fader I/Q I/Q With fading, power cal needs to be averaged over long periods of time Calibrate individual carrier Repeat for every frequency point Complete process could take hours or days Automated digital power cal completed in seconds One-button setup on the PXB calibrates power up to Page 30 15
Performance & Scalability to Meet Future Testing Needs Problem Solved: Engineers need an easy upgrade strategy to ensure their purchase solves today s requirements and tomorrow s Best way to meet emerging wireless requirements with higher performance Up to 120 MHz bandwidth 512 Msa playback memory Cost effective way to support higher order MIMO Buy what you need today Upgrade DSP blocks and I/O cards in one hour Automated self-test to verify installation Page 31 Performance & Scalability to Meet Future Testing Needs Problem Solved: Engineers can reuse leading edge Agilent instruments and applications Signal Inputs Signal Creation Tools Signal Outputs Analog I/Q - Direct from PXB - Connect to any DUT or vector signal generator with analog I/Q inputs Digital I/Q N5102A Page 32 MXA PXB ESG or MXG Page 32 16
and BB Design Integration Example Base Station Tx0 h00 h01 Rx0 Mobile BB and integrated sub-system Tx1 h10 Rx1 h11 PXB MXG/ESG Signal creation with Signal Studio, SystemVue 2008, ADS, or Matlab MXG or ESG for Page 33 Conclusion Wireless channels are complex environments that can hurt or help system performance MIMO and STC can greatly improve system performance Low channel correlations are required for MIMO Channel models need to include spatial correlations The PXB accelerates time to market with performance baseband Page 34 17
Additional Resources For more information, go to www.agilent.com/find/pxb Application note: MIMO channel modeling and emulation test challenges MIMO poster: 10 things you should know about MIMO Data sheet: Agilent N5106A PXB MIMO receiver tester Page 35 18