A Single RF Channel Smart Receiver Array with Digital

Darren Goshi, Yuanxun (Ethan) Wang, Tatsuo Itoh

University of California, Los Angeles Electrical Engineering Dept. Los Angeles, CA

University of California, Los Angeles

Single RF Channel Smart Antenna System

LPF ADC

LNA Mixer LPF ADC Digital MUX DEMUX Beam- Former

LPF ADC

Antenna Single RF Array Channel IF/Baseband Processing

¾MUX reduces N-channels (LNAs, Mixers) to 1 ¾Flexibility in IF/baseband processing – analog or digital DEMUX ¾Digital Beamforming in MATLAB environment

University of California, Los Angeles System Features

9 Reduction in costly RF hardware 9 Space reduction 9 Reduction in power dissipation 9 Maintains complete functionality as with typical smart antenna arrays

University of California, Los Angeles

Traditional DBF Smart Antenna System

LNAs Mixers

LPF ADC

LPF ADC Digital Signal Processing / Beamforming

LPF ADC

¾N-channel array requires N individual receiver chains ¾Each channel must be well matched and calibrated

University of California, Los Angeles System Principles

Switching waveform

Г=1 (when switch is open)

¾SMILE (Spatial MultIplexing of Local Elements) scheme 9 Sequential switching samples signal at each element 9 Sampling rate greater than Nyquist rate 9 Parallel feed network retains all signal information 9 Sampling frequency limits signal modulation bandwidth

University of California, Los Angeles

System Principles

Sampling Frequency:

fBNs ≥×

LPF Cutoff Frequency: BB <<−ff 22lpf s

University of California, Los Angeles System Principles

0.5

0

-0.5 18 19 20 21 22 23 24 25 0.5 o 0 ¾Baseband data recovery

-0.5 18 19 20 21 22 23 24 25 0.5

0 9 Non-conventional feed Array at 0 -0.5 18 19 20 21 22 23 24 25 0.5 network properly retains 0

-0.5 18 19 20 21 22 23 24 25 time µs phase information 0.5

0

-0.5 9 Allows use of advanced 18 19 20 21 22 23 24 25 0.5 o 0 beamforming and DOA -0.5 18 19 20 21 22 23 24 25 0.5

0

Array at 30 Array algorithms -0.5 18 19 20 21 22 23 24 25 0.5

0

-0.5 18 19 20 21 22 23 24 25 time µs

University of California, Los Angeles

4-Element Prototype Hardware

¾LNA MUX 9 NEC32485C 4-Element Pseudomorphic HJ FET (LNAs) as switching element MUX 9 Low-noise switching improves overall noise figure Feed Network with compared with passive Single RF Output switch 9 No current required to LNA MUX with patch drive switching antenna element at 5.8 GHz

University of California, Los Angeles Non-conventional Feed Network

Non-conventional array feed network

¾Always Matched Feed Network 9 Only one active channel at each instant 9 All lines at T-junction matched to Zo 9 No loss compared with 6 dB loss of Wilkinson dividers

University of California, Los Angeles

Circuit Schematic

From Antenna

Gate Bias LNA ¾Multiplexing Hardware Circuit NE32485C HJFET 9 Tank circuit adds isolation

sacrificed with non-conventional Tank Circuit feed network

9 Switch reference must be l λ/2 precisely matched to feed network Switch Zo Reference 3λ/2 Γ≈1 when off Zo

Z o To Single RF Channel

University of California, Los Angeles Single Source Testbed Setup

Antenna chamber SMILE array

LNA Horn antenna LO RF Gen

Digital PC Data oscilloscope acquisition

University of California, Los Angeles

Two Source Testbed Setup

RF Gen RF Gen

Horn antenna Horn antenna

SNOI SOI

SMILE array LNA

LO

Digital PC Data oscilloscope acquisition

University of California, Los Angeles Beamforming Performance

Synthesized Beam Patterns ¾Digital Beamforming 9 Synthesized beamforming LNA MUX Switching accomplished in DSP 0 0 -40 40 ¾FET Switching Benefits -5 9 Fast switching with FET -10 increases possible data rate

Amplitude [dB] Amplitude -15 9 Broader Symmetric Scan

-20 Range

-25 9 Lower sidelobes -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 Theta [deg]

University of California, Los Angeles

Digital Data Recovery

Demodulated Data 0.3

0.2

0.1

0

-0.1

-0.2

0 1 2 3 4 5 6 7 8 9 10

After Digital Beamforming

0.2

0.1

0

-0.1

-0.2

0 1 2 3 4 5 6 7 8 9 10 Time µs 200 kbps 600 kbps 1 Mbps

¾Digital Data

9 BPSK modulated carrier 9 Compare demodulation before and after applying DBF 9 Digital beamforming algorithm coherently sums all channels

University of California, Los Angeles DOA Estimation and Synthesized Patterns

Single Source 0 Synthesized Pattern DOA estimation

-5 ¾DOA Estimation

-10 9 Proper recovery of IF channel data

-15 allows advanced beamforming

-20 algorithms and DOA estimation

-25 9 MUSIC algorithm detects proper -100 -80 -60 -40 -20 0 20 40 60 80 100 Angle Two Sources direction of signals 0 ¾Synthesized Beamforming -5

-10 9 Beam formed indirection of source

-15 9 Null formed in direction of unwanted

-20 signal

-25

Synthesized Pattern DOA estimation -30 -100 -80 -60 -40 -20 0 20 40 60 80 100 Angle

University of California, Los Angeles

Spatial Filtering

Before Digital Beamforming 0.3

Single Source: 0.2

0.1

0

-0.1

-0.2

-0.3 0 5 10 15 20 25 30 35 40 45 50

After Digital Beamforming 0.3

0.2

0.1

0

-0.1

-0.2

-0.3 0 5 10 15 20 25 30 35 40 45 50 Time, us ¾Single Source Data Recovery

9 Coherent summation recovers a much clearer pattern 9 Results more evident for angles farther off broadside

University of California, Los Angeles Spatial Filtering

Before Digital Beamforming 0.3

Two Sources: 0.2

0.1

0

-0.1

-0.2

-0.3 0 5 10 15 20 25 30 35 40 45 50

After Digital Beamforming 0.3

0.2

0.1

0

-0.1

-0.2

-0.3 0 5 10 15 20 25 30 35 40 45 50 Time µs ¾Two Source Data Recovery

9 Incoherent demodulation does not resemble a digital pattern 9Interference ‘nulling’ demonstrates spatial filtering

University of California, Los Angeles

Conclusion

¾Measured 20 MHz switching rate ¾DOA Estimation and Beamforming ¾Recovered up to 1 Mbps digital data ¾Demonstrated full smart antenna functionality

University of California, Los Angeles Other Works

¾Slot Antenna Based SMILE Array 9 Compact design 9 Series fed PIN diode switching

¾2-D Configuration 9 Increase scanning flexibility for possible application

University of California, Los Angeles