UWB

1
UWB

• Ian ODonnell, Bob Brodersen

Berkeley Wireless Research Center Univ. of
California, Berkeley
2
The Known World
Traditional Radio Spectrum Allocation

• Narrowband
• (Bandwidth tightly constrained)
• Relatively small amount of unlicensed BW
• (Bandwidth scarcity)
• Spectrum Auctions ()
• Current spectrum poorly utilized

Radios Built/Researched To Those Specifications

• fcarrier, channel BW, adj. chan. interference,
  modulation, PTX, RX sensitivity (NF), transmit
  distance, etc.

3
Currently Available LAN/PAN Radios
Naïve Plot Transceiver Power vs. Throughput
Sources ISSCC 2001-2004 JSSC 1998 ISCAS
2000 RFM AD Honeywell Maxstream OKW IA(Ezrad
io) Chipcon Atmel 802.15.3a 802.15.4a Bluetoot
h, 802.11b, PicoRadio, c/o M. J. Ammer
UWB

• 1Mbps/
• 1mW
• (BIT)
• 110Mbps/
• 100mW
• (802.15.3a)

Throughput (bps)
TXRX Power (mW)
4
The Lure of UWB
Conventional Integrated Narrowband Transceiver
UWB Promises

• Simplicity
• Low Cost
• Integration
• Low Power
• Large BW
• Ranging
• Unlicensed Operation
• Coexistence

UWB Mostly Digital Radio
5
FCC Regulations

• Categorizes three types of UWB devices
• Imaging systems including ground-penetrating
  RADAR (GPR), through wall and medical imaging,
  and surveillance devices
• Vehicular RADAR systems
• Communication and measurement systems

• Specifies
• Minimum BW (-10dB) of 500MHz or 0.2fcarrier
• Peak to average ratio 20dB (max)
• Roughly two bands 0-960 MHz, 3.1-10.6 GHz
• 3m EIRP power levels (generally part 15)
• -20dBm total power in 0-960 MHz
• -2dBm total power in 3.1-10.6 GHz

But, Physical signaling not specified. Power
levels are low. (Shorter range)
6
UWB Terra Incognita
Bandwidth Limited
Energy Limited
UWB
Usual goal
Low signal to noise ratio Bandwidth
inefficient. (Here Thar Be Monsters!)
7
Standards 802.15.3a
Seems to be the main focus of the UWB in the
media (Hey, bandwidth is sexy, and who likes
wires anyways?) Purpose Small form-factor,
short range (due to limited TX power) but very
high rate communication. Ideally less
power/expense than WLAN (802.11a) Goal 110Mbps
at 10m in 100mW 200Mbps at 4m in 250mW
Optional 480Mbps at whatever distance can be
achieved Standard Stalemate Consumer
Electronic Companies (BOK) vs. PC Companies
(OFDM) Intels Wireless USB announcement may
trump standard
8
Standards 802.15.4a

• Low Rate Alternative PHY for WPAN
• Principal interest in ultra low power
  communication and precise ranging with scalable
  data rate/range.
• PHY Technical Requirements
• Data rate 1kbps up to 1Mbps (aggregators)
• Range 0-30m
• Power consumption several months to years
• Precision ranging (Location awareness 10s cm to
  1m)
• Dynamic networking
• Antenna non-directional
• Form factor appropriate for sensor network/RF
  tags
• Motion tracking (pedestrian/industrial vehicle,
  opt. automobile)
• Robust operation
• Call for Proposals Sept. 04 Done by Q2 2006

9
802.15.4a Applications

• Categories of Applications
• Safety/Health Monitoring
• Man down, search and rescue, situation
  awareness, etc.
• Personnel Security
• Tracking child/prisoner/guard, surveillance, auto
  car/workstation locks, etc.
• Logistics
• Tracking wildlife/cattle/workforce/customers/packa
  ges, space utilization measurements, robotic
  mowing/farming, call fwd, etc.
• Industrial Inventory Control
• Autonomous manifesting/meter reading, asset
  tracking, etc.
• Industrial Process Control and Maintenance
• Wireless sensor networks, robotics,
  anti-collision, remote sense/service
• Home Sensing, Control and Media Delivery
• Gaming, streaming media, media redirection, smart
  tags, HVAC control, etc.
• Communication
• PAN, Cordless phones, routing

(IEEE P802.15-03-04420-01-004a)
Many disparate applications/markets.
10
Why Use 0-1GHz?
Considered to be crowded or dirty spectrum
(Due to the historical development of wireless,
it is the heaviest used part of the spectrum.)

• Why is it desirable?
• Good material penetration
• Longer transmit distances
• Low frequency is easier for design
• Lower power consumption (lower frequency
  operation)
• What is wrong with it?
• Lower frequencies mean larger passive values for
  filters, harder to integrate
• Larger antennas (related to l)
• That aforementioned interference problem from
  preexisting users

Fig. c/o Bob Scholtz
11
Spectrum Usage 0-1GHz
Some Interferers FM 88-108 MHz TV 54-88 MHz
(VHF 2-6) 174-216MHz (VHF 7-13)
470-806MHz (UHF 14-69) Taxi 157, 452,
457MHz Police 154-6, 158-9, 460, 465
MHz GMRS 462, 467MHz ISM 902-928MHz Cell
phone824-849MHz, 870-893MHz Pager
929-930MHz
Maybe not as bad as expected
12
Pulse-Based Throughput of 0-1GHz
Using previous interference, capacity is
28.5Gb/s! (3m) Even with noise equal to the
mask, there is 2Gb/s (3m) If you use a pulse,
an optimistic link budget (using prev.
interference) yields 100Mbit at 3m 1Mbit at
60m FCC Mask Part 15.209 (Class C Emission)
for low-frequency imaging systems
13
Why Ranging Is For Free

• Radios need to be synchronized to communicate.
• Synchronization requires fine timing resolution
  for UWB pulses.
• That timing resolution (500ps for 2GSa/s) may be
  easily used to measure time-of-flight. (0.5 ft)

Can improve time-of-flight measurement by
examining correlation profile, allowing for
sub-500ps resolution (on the order of inches)
with 0-1GHz.
14
The UWB Opportunity

• What is hard about UWB?
• Antenna (omni, small form-factor, efficient
  coupling)
• Interference (filtering, tolerating in-band
  interferers)
• Spectral mask compliance (pulse/waveform
  generation)
• Fast acquisition, Efficient synchronization
• Efficient channel estimation
• Low power consumption w/ 1GHz BW and 2GSa/s ADCs
• Exploiting range information (networking)
• System exploration space is large

But for researchers, this is exciting stuff !
15
Research Areas

• UWB Channel/Interference Meas./Modeling
• Antenna Research
• Circuit Research
• Antenna Co-Design
• Pulse generation
• System Design
• Pulse-based
• OFDM
• Transmit-reference
• 1-bit mostly-digital
• Signal Processing/Algorithms
• Rapid Acquisition, Synch, Locationing, Imaging
• CAD Tools (Simulation and Design)

• Min. power consumption
• Min. jitter
• Frequency-based (bandpass)
• Orthogonal projection
• Subsampling
• Oversampling

16
UWB Channel Research/Measurements

• Spatial Capacity
• Recent Ph.D. Thesis
• (Ada Poon)
• Spatial Channel Measurements
• (Jing Yang)
• Sweep TX/RX antenna arrays over 360o azimuth,
  -40o to 90o elevation
• TX from pulse generator
• 20GSa/sec Scope as RX

17
Small Antenna Modeling

• Desirable UWB antenna characteristics
• Small form-factor
• Broadband
• Omni-directional

6cm Dipole Antenna Input Impedance

• For 0-1GHz, small is electrically small
• E-fields nearly same in all dir
• Simple circuit model
• Can estimate radiated E-field in SPICE
  (Co-Design)
• But, antenna is inherently narrowband
• Inefficient radiation
• 50 W radiation resistance not possible over whole
  BW

c/o Stanley Wang
18
Antenna/Circuit Interface
STMicroelectronics 0.13um CMOS process

• Transmitter Pulse generation
• Total area 0.49mm2
• 1.2V Vdd
• Differential drivers
• 16 levels of drive

• Receiver LNA
• Current-reuse allows for 50 W Zin with 1mW
  consumption
• Layout area 59mm x 45mm
• Coupling capacitors off-chip

c/o Stanley Wang
19
Berkeley Impulse Transceiver (BIT)
Targeting Sensor Network Application

• Specifications
• 100kbps over 10m with 10-3 BER
• 1mW total (TXRX) power consumption
• 0-1GHz bandwidth

GAIN
ADC
CLK
DIGITAL
TX
First all-CMOS integrated UWB transceiver for
comm. and ranging/locationing Aggressive
low-power design Mostly-digital approach,
simplify analog front-end Provide flexible
platform for further research
20
Pulse Reception
Only process data from a window of time
Voltage
time
Sample Time
Pulse Reception Window
Pulse Transmission Rate
Analog On Sampling On Digital Off
Analog Off Sampling Off Digital On
Analog Off Sampling Off Digital Off
Analog On Sampling On Digital Off
Receiver Operation
time
21
Transceiver Power Estimate
Always On 8 mW (32 Mpulse/s)
Duty-Cycling Starts
GAIN
DIGITAL
BIAS
A/D
OSC
TX
DLL
CONTROL
Duty-Cycled To 1mW (1 Mpulse/s)
22
Transceiver Status
O.K., this time for sure Chip will
tape-out this summer! Gain stages layout left,
and TX block, then final assembly.
Process 0.13um ST Microelectronics Size
15mm2 Digital 3.3mm x 3.3mm 245,000
StdCells Analog 3.3mm x 1.1mm
23
System Exploration/Prototyping
Berkeley Impulse Transceiver (BIT) (0-1GHz)
Also designed as flexible testbed
Different antennas/LNA impedance Variable
transmit power Variable pulse rate Programmable
pulse matched filter
Adjustable synch/data recovery Independent
synch/data PN sequences I/O can be sent from A/D
directly to BEE for more sophisticated signal
processing
Generic UWB Receiver (0-1GHz) Allow for network
exploration/algorithm development. 8-bit,
2GSamples/s. Data is de-muxed into 64 250MHz
streams then optically serialized for transmit to
BEE riser card. (Stanley Wang, Kathy Lu)
Our Next Generation UWB PHY (3-10GHz)
Explore architectures/trade-offs for low-power,
3-10GHz implementations. (Mike Chen)
24
Next Generation Networking

• How to connect potentially billions of low cost,
  low power, sensor/actuators/smart-devices?
• Want wireless for mobility, ease of deployment,
  low cost
• (e.g. wiring cost dominates for temperature
  sensors with HVAC)
• Issues
• Interference/coexistence (spectrum sharing)
• Access control
• Energy efficient routing/relaying
  (location-based routing)
• Scalability
• How best use position information
• Robust operation

25
Conclusion

• There are a wide variety of unique applications
  of this technology and a large amount of
  interest.
• UWB signaling provides a new way to utilize the
  spectrum and there is a wealth of possibilities
  for new system designs.
• Impulse-based signaling looks promising to
  realize novel, very low-power, moderate-rate
  ranging and communication radios.
• Current research is just the tip of the iceberg.
• Big opportunity for someone willing to try
  something new!

26
Acknowledgement

• This research was supported by
• The Office of Naval Research (Award No.
  N00014-00-1-0223),
• A Multidisciplinary University Research
  Initiative (MURI) grant from the Department of
  Defense (Grant 065861),
• National Science Foundation Next Generation
  Networking
• (Grant ANI0230963)
• And the industrial members of the Berkeley
  Wireless Research Center.
• Also, thanks to
• ST Microelectronics for access to their 0.13mm
  process,
• BWRC students M. Josie Ammer, Ada Poon, Jing
  Yang, Stanley Wang, Mike Chen, Kathy Lu for
  their input and slide fodder.
• Aether Wire Location for inspiring us to start
  investigating impulse radio.
• Maps from www.cosmography.com and
  infocom.elsewhere.org