Low Rate UWB for TG4a

1
Project IEEE P802.15 Working Group for Wireless
Personal Area Networks (WPANs) Submission Title
TG4a Proposal for Low Rate DS-UWB
(DS-UWB-LR) Date Submitted January
2004 Source Matt Welborn (1), Michael
McLaughlin (2), Shariar Emami (1) Company
(1) Freescale Semiconductor, Inc, (2) decaWave,
Ltd. Address 8133 Leesburg Pike, Vienna VA
22182 Voice703-269-3000, FAX ,
E-Mailmatt.welborn _at_ freescale.com, or
michael_at_decawave.com Re Response to Call for
Proposals Abstract This document describes a
proposal for the TG4a baseline draft
standard. Purpose Proposal Presentation for
the IEEE802.15.4a standard. Notice This
document has been prepared to assist the IEEE
P802.15. It is offered as a basis for discussion
and is not binding on the contributing
individual(s) or organization(s). The material in
this document is subject to change in form and
content after further study. The contributor(s)
reserve(s) the right to add, amend or withdraw
material contained herein. Release The
contributor acknowledges and accepts that this
contribution becomes the property of IEEE and may
be made publicly available by P802.15.
2
UWB for Low Rate Communications

• UWB has great potential for low power
  communications
• Low fading margin can provide same range for
  lower transmit power
• Large (ultra-wide) bandwidth can provide fine
  time resolution provides potential for accurate
  ranging primary differentiation from existing
  15.4 capabilities
• Drawbacks due to regulations
• Limited transmit power how much is enough?
• Operation at long ranges is highly dependent on
  NLOS path loss characteristics

3
Proposal for TG4a Alternate PHY Layer
4
Overview of DS-UWB-LR TG4a Proposal

• Based on higher rate DS-UWB proposal under
  consideration in TG3a, but modified for low power
  and lower complexity
• Variable-length code spreading with BPSK
  modulation of data
• Chip rates are 1/3 the rates used in DS-UWB (so
  about 450 MHz)
• Bandwidth is only 500 MHz (instead of 1500 M)
• Lower complexity FEC convolutional code
  (constraint length k4)
• Data rates of 30 kbps to 10 Mbps
• Use spreading codes of length 24 to 6144
• AWGN range of 15 to 75 meters (assuming n3.5 PL
  at gt 10m)
• Support for precise ranging
• Very straight forward solutions using RTT
  measurements with corrections for non-direct-path
  propagation effects
• Flexible pulse shaping
• Allows many pulse generation, antenna and
  receiver architectures
• Supports requirements for coexistence
  regulatory constraints
• Will enable interoperation between higher rate
  DS-UWB devices and low rate, low complexity TG4a
  devices
• Support wider range of asymmetric applications
• Enables active coexistence and coordination

5
Benefits of Low Rate DS-UWB

• Bandwidth operating frequency
• Transmit power, ranging, complexity performance
• Pulse rate
• Effects on efficiency implementation
• Data Rate
• Interoperability Coexistence

6
Operating Frequency

• If multiple bands used
• Multiple operating channels with different center
  frequencies to provide FDM operation for
  different networks
• Three piconet bands at 3536, 4056, and 4576 MHz
• Different performance due to 20 Log10(Fc) term in
  path loss
• Also uses CDM for sharing of each band by
  different piconets
• Could also use pure CDM in a single wider band
• Could even mix wider band devices (1500 MHz) and
  narrower band devices (500 MHz)
• Cost of generating the reference frequency
  depends on the specific frequency
• DS-UWB is based on low cost, high quality 26 MHz
  crystals (widely used in cell phones)
• Better frequency accuracy can relax other system
  constraints
• Acquisition at longer range requires longer
  integration and therefore more accurate reference
  frequency
• High accuracy clock can allow longer sleep time
  better power management
• Precise ranging using TOA methods requires high
  precision time measurements over relatively long
  intervals (RTT) to determine small differences in
  signal propagation times
• Simplified and improved with good reference clock

7
Signal Operating Bands for Low Rate UWB
Relative PSD (dB)
Possible Lower Rate Signaling Bands (500 MHz
bandwidth)
High rate DS-UWB Low Band with RRC Pulse Shape
0
-3
FCC Mask
-20
4056
3536
4576
Frequency (MHz)
3100
5100

• Ultra-wideband 500 MHz bands for each piconet
• Code-division and Frequency-division multiplexing
  
• Multiple piconets in each band using different
  codes
• Operation in close proximity, interference
  avoidance or coexistence
• Three piconet bands at 3536, 4056, and 4576 MHz

8
Low Rate DS-UWB Pulse Rate

• Impulse radio (IR) originally meant low pulse
  rate (10s of M pulse/sec) using time hopping
  for multiple access and pulse position modulation
  (PPM)
• More generally, IR is just pulse-based spread
  spectrum with data modulation
• Many choices for modulation (BPSK, PPM, OOK,
  etc.)
• One or more pulses per data symbol
• Low rate DS-UWB uses a chip rate that is designed
  to meet minimum 500 MHz bandwidth for simple BSK
  modulation
• Center frequency is always a multiple of 26 MHz
• Chip rate is equal to center frequency divided by
  9 (e.g. 4052 / 9 450.66MHz)
• Spreading codes are based on 24-chip code
• Longer spreading codes are derived from the basic
  code by further sreading with PN sequence (e.g.
  length 192 code is a length 24 code spread by a
  length 8 PN code)
• Pulse rate does not fundamentally affect transmit
  power, signal bandwidth or system performance
• Pulse rate does affect energy per pulse and
  therefore peak power (and voltage)

9
Higher Pulse Rate Lower Peak Power
Higher peak power voltage for same average
power

• Lower pulse rate requires higher energy per
  pulse and therefore higher peak power (and
  voltage) for same transmit power
• Process technology can limit available peak
  voltage that can be achieved without an external
  power amplifier ___
• If pulse rate is 100x slower, then peak voltage
  is ?100 10x higher

10
Something Special about Impulse Radio

• For impulse radio, basic idea was to send a fixed
  pulse shape at a regular (or controlled) interval
• For conventional carrier systems, the pulse
  shape at RF is not fixed due to lack of fixed
  relationship between carrier phase and symbol
  clock phase
• Result if properly designed we can have combine
  the benefits of conventional carrier-based
  systems and impulse systems
• Fixed phase relationship between carrier symbol
• DS-UWB is based on integer relationship between
  center frequency (carrier) and chip or symbol
  rate
• Makes frequency synthesis and implementation
  easier

11
No Advantage to Non-Coherent
Given that synchronization to chipping clock must
be done regardless, and Since RF is easily
synchronized to the chipping clock, There is no
reason not to build coherent systems
Chipping Clock Must Be Synchronized Regardless of
Coherent/Incoherent Approach
-
Ring
-
up a filter to create
an RF UWB pulse

Simple injection lock or PLL
Ring
No jitter issues because of high BW
Oscillator
(only 3X multiply)
A very small ring oscillator at 4.1 GHz

Injection lock or PLL
-
an LC oscillator
12
Data Rate Considerations

• Lowest PHY data rate does not necessarily mean
  lowest energy consumption
• In fact, a fast radio can potentially be more
  energy efficient than a slow radio. Example
• Compare a 1 Mbps radio at 100 mW versus 10 kbps
  radio at 10 mW
• 32 kB _at_ 10 kbps 0.256 mWseconds
• 32 kB _at_ 1 Mbps 0.0256 mWseconds 1/10 of the
  energy per bit!
• Assumptions
• Both radios achieve minimum range requirement for
  application
• Minimum acquisition time is a function of SNR
  (range) not data rate
• Requires fast wake-up and shut down of radio with
  aggressive power management
• Relative energy usage depends on packet size
• Fast radio advantage is higher for longer packets
  
• Notice transmit power is a small fraction of the
  total power (lt1)
• The largest power use is turning on the radio and
  processing signal

13
Data Rate Considerations
Lower radio power, but longer transmission time
for data payload
Power
Preamble
Time
Higher peak power but shorter transmission time
for same payload
Power
Preamble
Time

• Total energy use from battery is the area under
  the power vs. time curves shown above
• Relative efficiency depends on power duration
  (payload size)

14
Low Rate DS-UWB Data Rates

• LR-DS-UWB data rates are designed to support
  relatively high rate operation at relatively
  small duty cycle

Spreading code Length Symbol Rate FEC Rate PHY Bit Rate
192 2.35 MHz 0.5 1.17 Mbps
768 587 kHz 0.5 293 kbps
6144 73 kHz 0.5 37 kbps
24 18.778 MHz 0.5 9.4 Mbps
15
Link Budget
16
Link Budget Notes

• Assumptions
• Chipping rate 450.66 MHz
• Includes 1.9 dB Tx power reduction for spectral
  ripple in certification testing
• Assumes 3.8e-5 BER to achieve 1 PER with 32
  octet data packet

17
Interoperability Coexistence

• All of the specific co-existing systems in the
  Slection Requirements are out-of-band to
  LR-DS-UWB
• Robustness against in-band interference is
  provided by the UWB bandwidth of the LR-DS-UWB
  system (large processing gain)
• Many types of other UWB systems and waveforms
  will share the UWB bands
• Interoperability between TG4a higher rate
  systems could enable improved coexistence
• Interoperation with higher rate systems could
  increase the utility of the TG4a standard
• Interoperability of low cost sensor/RFID devices
  with nearby UWB CE devices
• Interoperability with DS-UWB could be quite
  simple if correct parameters are chose for TG4a
• Common reference frequency, codes operating
  bands

18
Dual 15.4 and DS-UWB (15.3a Proposal) is Easy and
Could Provide Interoperability/Co-existence
Receiver
LNA
ò
To ADC
S
/
H
Ternary Code
3
n RF Cycles


1
,
0
,
-
1

Encoded
Ant
per Chip
TxData
Code Gen
Filter
LPF
Code Clock
Chip Clock
d
Sense
r
D
x
3
n
Q
Ck
This is as simple as anything else
Symbol Rate
?L
code
Chip Rate Clock
Idealized
n

1
for
1
.
5
GHz BW
(
15
.
3
a
)
?n
-
-
-
1
1
1
1
0
1
1
1
n

3
for
500
MHz BW
(
15
.
4
a
)
Filtered (LPF RF)
1
.
352
GHz
(
104
13
MHz
)
Inverted
,
Non
-
Inverted and Zero
-
Amplitude Wavelets
Agile Clock
Locks to Chip Rate
19
Ranging using the DS-UWB Approach

• TG4a has done an extensive review of potential
  location techniques thank you!
• Some result in hyperbolic lines of position (LOP)
• Some result in circular LOPs (since they are
  determined directly from node-to-node ranges)
• Simple node-to-node ranging can also be useful
  when used without the benefit of fixed reference
  nodes
• Brief review of node-to-node ranging

20
Basic Approach for Ranging
Packet Exchange Timeline
Delay before data is transmitted
Total elapsed time
Response
Initiating packet
Node A
Node B
Initiating packet
Response
Delay in receiver before ACK is transmitted
?T due to multipath excess delay
?T due to direct path propagation
21
Required Measurements
Node A Computes Range
Node A Initiates Ranging Operation
TX_Initialize (Node A)
Sync_Ref_Point (Node A)
Node A Excess Delay Estimate
Total elapsed time
Node A
Channel Scan
Response packet
Initiating packet
Node B later sends data to A
Sync_Ref_Point (Node B)
TX_Initialize (Node B)
Inter-packet time
Node B
Initiating packet
Response packet
Channel Scan
Node B Excess Delay Estimate
22
Example NLOS Range Calculations
Actual Ranges (ft) Actual Ranges (ft)
1 16.7
2 17.7
3 18.6
4 19.6
5 20.6
6 21.6
7 22.5
8 23.5
9 24.5
Nine Different NLOS Channel Impulse Responses,
Ranges 16-25 Feet
-3
x 10
3
2
Amplitude
1
0
-1
-2
-3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
-7
x 10
Time (seconds)
Leading Edges of Reponses, Showing Direct Path
Component Detection
-3
x 10
1.5
16.7 ft-1-1
17.9 ft-2-1
24.1 ft-4-1
18.7 ft-3-1
1
Amplitude
0.5
0
-0.5
20.5 ft-5-1
23.5 ft-8-1
22.4 ft-7-1
-1
21.4 ft-6-1
21.2 ft-9-1
-1.5

200
220
240
260
280
300
320
Range (in inches)
23
Measured Multipath Resolution with an Operating
Radio
Optimal lock path for radio
14 dB lower power
Earliest arriving path
Multipath component amplitude (dB)
10 ns earlier
Time in nanoseconds (time reversed)
24
Technical Feasibility

• The Low rate DS-UWB solution for TG3a has a high
  level of manufacturability
• Based on existing DS-UWB technology
• Can be implemented in low-cost CMOS technology
• Time to Market
• Time to market should be quite reasonable.
• Regulatory Impact
• DS-UWB technology is know to be compliant with
  FCC UWB rules
• Other regulatory administrations are using FCC
  rules as a basis for initial discussions
• Many mechanisms exist to ensure compliance for
  other regions that adopt other regulations

25
System Performance

• System and SOP simulations are underway results
  TBD

26
Collaboration Whats Important
Characteristic Importance (to me)
Radio manufacturability High
Performance-to-emissions ratio High-to-medium
Absolute performance Medium
Precision ranging potential High-to-medium
Coexistence w/other UWB High-to-medium
Regulatory compliance Essential
Scalability (rates, applications, etc) High-to-medium
Ultra-low complexity Medium
Low power consumption support High-to-medium
Signal bandwidth (500 to 1500) Medium-to-low
Other modulation parameters Specific spreading codes, chip rates, data rates, etc Medium