Pattern-Reconfigurable Antennas Optimized for Automotive Applications CST European Automotive Workshop, 23.11.2015 Jerzy Kowalewski, Tobias Mahler, Thomas Zwick INSTITUT FÜR HOCHFREQUENZTECHNIK UND ELEKTRONIK National Research Center of the Helmholtz Association www.ihe.kit.edu Introduction Multiple input, multiple output systems (MIMO) for automotive wireless communications Multiple antennas with one frontend per antenna Use of two effects: Parallel transmission in space, multiplex (high SNR) Diversity, Beamforming (low SNR) Goals: increase of data rate and reduction of outage probability MIMO channel Source: http://blog.nxp.com/wpcontent/up loads/2013/12/car-audio- Entertainment. jpg (13.07.2015) 2 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
MIMO: Resulting Beam-forming and Multiplexing Space-Time beam-forming a 4 a 3 a 2 a 4 a 1 a 3 a 2 a 1 MIMO processing 4 dipole lin. array, /2 spacing building propagation path 1. sub channel 2. sub channel with channel state information Simulations based on IHE ray tracing 3 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Important Considerations A MIMO system requires a full transceiver for each Tx and Rx antenna and a powerful digital signal processing. Most of the time the number of sub-channels used is smaller than the number of Tx/Rx antennas Typical MIMO systems use identical, omni-directional antennas Is there a better solution for the automotive case? Reconfigurable MIMO antenna: less antennas and transceivers but beam switching for each antenna 4 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
MIMO Measurement Setup 4x4 MIMO measurement system MIMO channel measurement and capacity estimation Measurements in urban environment 5 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler First Measurement Results A SNR = 30dB SNR = 10dB 6 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
First Measurement Results B Eigenvalues of mobile channels Channel capacity for 10% outage probability 7 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Synthesized Optimal Patterns Urban Pattern 1 Urban Pattern 2 Driving direction Target patterns to be realized by antenna designer 8 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Antenna Principle 1 Series Feed Antenna mounted vertically on the roof (56 50 cm²) Two monopoles separated by Driving direction Series feeding Two lines from left to right monopole (length λ) 0 or 180 phase shift depending on switch settings Distance between junction and switches - about 9 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Simulation Results (Serial Feed) Off On State 1 G = 7.5 dbi G orth = 6.6 db State 1 (Surface currents) On Off Gain dbi State 2 G = 6.3 dbi G orth = 12 db State 2 Driving direction 10 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Antenna Principle 2 Parallel Feed Antenna mounted vertically on the roof (50 50 cm²) Two inverted L antennas separated by Parallel feeding Line from left to right monopole (length λ) T junction in the middle Microstrip balun 0 or 180 phase shift depending on switch settings Driving direction 11 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Switch Configuration (Parallel Feed) DC pin State Switch 1 (front up) Switch 2 (front low) Switch 3 (back) V DD Port DC 1 Off On On +10V +1V 2 On Off Off -10V 0V 12 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Simulation Results (Parallel Feed) Gain dbi State 1 (Surface currents) Driving direction State 1 G = 6.5 dbi G orth = 18.9 db State 2 State 2 G = 5.6 dbi G orth = 4 db 13 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler HF-Switches Pin-diodes Forward current: On Reverse voltage: Off Power dissipation: mw Non-linear device MEMS-switch Micro-electro-mechanical system Electrostatic: high voltage Power dissipation: μw High isolation Comparable insertion loss Good models needed for accurate modeling 14 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Measurement Results of MEMS and Pin Diode 0 MEMS 0 MEMS Insertion loss in db -1-2 -3 Datasheet Messung Measurement Datenblatt -4 0 2 4 6 8 Frequenz Frequencyin in GHz Isolation in db -1-20 -2-40 -3-60 -4 0 2 4 6 8 Frequency Frequenz in in GHz 2,45 GHz MEMS Pin-Diode (NXP BAP64-02) Meas. Datasheet Meas. (as in datasheet) Insertion loss 0,45 db* 0,3 db 0,35 db* (10 ma) Isolation 15,6 db 30 db 13 db (10 V) * Attenuation of transmission line considered 15 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Switch Model Simulations in CST with equivalent circuit model Pin-diode modeled with 10 ma bias (MEMS similar) 0-5 Einfügedämpfung Insertion loss in in db db -0.5-1 -1.5-2 -2.5 pin CST MEMS 1 2 3 4 Frequency Frequenz in in GHz GHz Isolation in db db -10-15 -20-25 -30 1 2 3 4 Frequency Frequenz in GHz 16 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
DC Bias Realization (Serial Feed Pin) Pin-diode Bias-Tee Diodes antiparallel mounted Resistor shorted to ground ±2 ma SMA-port Front view (pin) 17 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler DC Bias Realization (Serial Feed MEMS) MEMS-switch DC lines on antenna substrate Resistors shorted to ground Parallel to right monopole Resistors in DC-lines used as RF Chokes DC-pins underneath the ground Negligible influence SMA-port Front view (MEMS) SMA-port Back view (MEMS) 18 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Antenna fabrication Rogers RT5880 substrate 0.8 mm ε r = 2.2 FR4 as ground plane board Easy soldering pin MEMS Copper foil on Styrodur as roof 19 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Measured Matching Shift of the resonance frequency of 50 MHz and 100 MHz Switch model causes bigger phase shift than expected 10 db-bandwidth (6 db-bandwidth) pin: 120 MHz (240 MHz) MEMS: 140 MHz (270 MHz) pin-diode MEMS-switch 20 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Measured Radiation Pattern for State 1 Azimuth 0 Meas. pin Sim. pin Meas. MEMS Sim. MEMS Azimuth 270 Measurement at 2.45 GHz Gain 4.5 dbi and 5.5 dbi respectively G orth 9 db bzw. 17 db 21 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Measured Radiation Pattern for State 2 0 0 45-45 45-45 90-20 -10 0 10-90 90-5 0 5 10-90 135 180 Azimuth 0 dbi -135 Meas. pin Sim. pin Meas. MEMS Sim. MEMS 135 180 Azimuth 270 dbi -135 Measurement at 2.45 GHz Gain 4 dbi and 3 dbi respectively G orth 11 db bzw. 9.5 db 22 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Measured Radiation Pattern for State 2 0 0 45-45 45-45 90 135-90 -20-10 0 10 dbi -135 180 Azimuth 0 Meas. pin Sim. pin Meas. MEMS Sim. MEMS 90 2.32 GHz 2.36 GHz 2.44 GHz 135 2.48 GHz 180 MEMS: Azimuth 270 0 3 6 dbi -135-90 Lower gain at target frequency due to frequency shift 23 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Conclusions Multiple input, multiple output systems (MIMO) for automotive wireless communications promise an increase of data rate and a reduction of outage probability Reconfigurable MIMO antennas with less antennas and transceivers but beam switching for each antenna could be a better compromise between cost and performance especially regarding the reduction of outage probability 24 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler
Danke für Ihre Aufmerksamkeit! 25 26.11.2015 Thomas Zwick, Jerzy Kowalewski, Tobias Mahler Multiple antenna systems Multiple input, multiple output (MIMO) Cost on the mobile user side grow with number of antennas/frontends Cost intensive processing Only two to three usable sub-channels due to simulation results MIMO processing 26 26.11.2015 Thomas Zwick, Jerzy Kowalewski
Reconfigurable (multiple) antenna systems Reconfigurable (multiple) antenna systems Choice of best directivity patterns Reduced number of frontends cost reduction Comparable performance MIMO processing 27 26.11.2015 Thomas Zwick, Jerzy Kowalewski Capacity of reconfigurable (multiple)antenna systems Number of selectable radiation patterns: N r Number of parallel (sub-)channels or patterns the antenna can handle simultaneously: N rp Single reconfigurable antenna with N rp =1 and N r =2 Multiple reconfigurable antenna with N rp =2 and N r =4 Maximum search over all possible switching combinations for each snapshot Example for Nrp=2 and Nr=3 with a single Tx antenna (2 of 3 receive modes are active) Number of possible switching combinations (with N r N rp ) 28 26.11.2015 Thomas Zwick, Jerzy Kowalewski
Simulation setup Exemplary urban RayTracing scenario Base Station (transmitter) Automotive pattern reconfigurable rooftop antenna (receiver) Reconfigurable Antenna 29 26.11.2015 Thomas Zwick, Jerzy Kowalewski Simulation results Urban scenario Rural scenario Relative Switching Occurrence Mode 1 Mode 2 Mode 3 Relative Switching Occurrence Mode 1 Mode 2 Mode 3 Utilization (single reconf.) 13,5 % 45,9 % 40,6 % Utilization (single reconf.) 0,2 % 67,7 % 32,1 % Non-Utilization (dual reconf.) 22,1 % 33,5 % 44,4 % Non-Utilization (dual reconf.) 20,5 % 20,7 % 58,8 % 30 26.11.2015 Thomas Zwick, Jerzy Kowalewski
Synthesized optimal patterns Urban Subchannel 1 Urban Subchannel 2 Rural Subchannel 1 Rural Subchannel 2 Driving direction 31 26.11.2015 Jerzy Kowalewski MIMO configuration Antennas mounted on an additional ground plane (60 94 mm²) Elements separated by 34mm Decoupling structure Slots etched in the ground Better isolation between ports Four patterns realizable by state switching of two antennas 32 26.11.2015 Thomas Zwick, Jerzy Kowalewski
Simulation results Driving direction State 11 State 12 State 11 G = 7.5 dbi G orth = 18 db State 12 G = 7.1 dbi G orth = 17 db State 21 G = 7 dbi G orth = 17 db State 21 State 22 State 22 omnidirectional G = 4 dbi 33 26.11.2015 Thomas Zwick, Jerzy Kowalewski Capacity: SISO versus MIMO without CSI SISO: Classical Shannon formula MIMO: Claude Elwood Shannon M C log 2 (1 SNR H 2 ) C log 2 1 P Txi 2 i i1 3 db increase in Signal to Noise Ratio gives another bit/s/hz efficiency For M Tx = N Rx = M antennas capacity grows linearly with large M SNR Signal to Noise Ratio; H complex transmission coefficient; * compl. conjugate transposed 34 26.11.2015
MIMO Capacity with CSI M=1 M=1 N=1 N=3 C log 2 (1 P Tx a ) log 2 2 TxRx 2 1 SNR h C log 2 det I N SNR hh N C CSI @Tx&Rx log 2 1 P Tx,i with channel 2 i state information i1 (CSI) at Tx summation of the sub-channel capacities C CSI @Rx log 2 det I N SNR M H without channel NxMH MxN state information (CSI) at Tx P Txi i 2 sub-channel i transmit power sub-channel Eigen value noise power h h H = channel impulse response = channel impulse response vector = MIMO channel matrix H + = hermitian matrix 35 26.11.2015 Simulated matching S 11 in db 0-10 -20-30 pin: pinzustand State 1 pin: pinzustand State 23 MEMS: Zustand State 1 MEMS: Zustand State 32-40 1,5 2 2,5 3 3,5 Frequenz in GHz State Switch 1 Switch 2 (upper) (lower) 1 Off On 2 On Off 3 Off Off Three usable states Resonances for different states slightly shifted MEMS-switches better Different DC-lines 10 db-bandwidth Pin: 130 MHz MEMS: 150 MHz 6 db-bandwidth Pin: 260 MHz MEMS: 300 MHz 36 26.11.2015 Thomas Zwick, Jerzy Kowalewski
Simulated matching S11 in db 0-5 -10-15 State 1 State 2-20 2 2.2 2.4 2.6 2.8 3 Frequency in GHz Two usable states Resonances for different states show no shift 10 db-bandwidth Pin: 110 MHz 6 db-bandwidth Pin: 250 MHz State Switch 1 (front up) Switch 2 (front low) 1 Off On On 2 On Off Off Switch 3 (back) 37 26.11.2015 Thomas Zwick, Jerzy Kowalewski