ATSC 3.0 Mobile Support

Luke Fay

Outline • ‘Mobile Service’ definition

– LTE currently is used by mobile [cellular] operators to deliver mobile service. But let’s look at what that means for broadcasters.

• Network support that works with LTE (use in broadcaster context) – Dense cell [yes] – Sparse cell [no] – Single large cell (big stick) [no]

• Broadcaster mobile desires – Definition of ‘mobile service’ for broadcasters – How proposed ATSC 3.0 PHY supports these needs

• Support for other standards by FEF (LTE, WiFi, others) • Summary

LTE ‘Mobile Service’ • Wireless communication of high-speed data for mobile

phones and data terminals.

• LTE’s ‘Lightning-Fast’ speeds are slower on average

than broadcast speeds of ATSC 1.0 today.

Network Downlink Uplink LTE access likelihood *

AT&T 18.6Mbps avg. 9.0Mbps avg. 81.7%

Sprint 10.3Mbps avg. 4.4Mbps avg. 50.2%

T-Mobile (non-LTE)

7.3Mbps avg. 1.5Mbps avg. N/A

Verizon 14.3Mbps avg. 8.5Mbps avg. 93.2%

Source: http://www.rootmetrics.com/us/blog/special-reports/lightning-fast-data-speeds-and-expanding-coverage

LTE is made for unicast and updated for multicast while keeping infrastructure for unicast.

*likelihood of accessing 4G speeds, connection speed drops to 3G rates if access is not acquired.

4

New EBU Technical Report 27 The EBU has released a report on “ DELIVERY OF BROADCAST

CONTENT OVER LTE NETWORKS” • Available at https://tech.ebu.ch/docs/techreports/tr027.pdf

Summary • The participant included broadcasters and mobile equipment

manufacturers but not mobile operators. • Technically free-to-air delivery “could in principle be enabled by LTE

eMBMS” • Mobile industry: Delivery costs for broadcast over LTE may be reduced

by an efficient combination of unicast and eMBMS capabilities • LTE delivery costs still seen as significantly higher than the current costs

of TV distribution • LTE for large-scale distribution of TV services and as possible

replacement for current DTV delivery are not clear and require further study LTE networks for a large scale TV distribution is not

envisaged in the short term

Network Support: Dense cell • 3GPP TS 36.201 V8.0.0 (2007-09) , TS 36.211 & TS 36.212 physical layer description

– 512 point FFT for 5MHz bandwidth (15kHz subcarrier spacing) over compensates for Doppler spread • Reference [1] shows that error-free mobile reception can be done with speeds up to 130km/hr with 500Hz

(MUCH LESS) carrier spacing at 600MHz RF center frequency

– Cyclic Prefix = 16.67 usec or 33.33 usec for eMBMS, no wide range of choices • Optimized for small cell network with unicast data • Receiver device battery power savings in mind (lower transmit power needed as towers are close together) • Covers multipath propagation within 5km, broadcast SFNs cover much larger areas resulting in longer delay

spreads. • Fixed values limits PHY flexibility, precludes sparse SFN

– FEC: 1/3 turbo code selected some time ago and likely stayed for legacy reasons • Unicast has TCP/IP communication where return link delay is of concern in QoS provisions for a transfer latency

of < 5msec in the Radio Access Network. • Broadcast is multicast data where TCP/IP links do not exist; UDP links exist and higher block lengths can be

used.

– eMBMS has to suffer all the choices of unicast LTE • Received signal strength is driving factor for mobile support

– Indoor reception on smartphones with antenna LOSS is compensated with many low power low tower cells

[1] Projektrat DVB-T2 Modellversuch Norddeutschland: Terrestrik der Zukunft – Zukunft der Terrestrik. Shaker Verlag. Aachen, 2012

LTE is optimized for unicast communications with dense cell networks.

Network Support: Sparse cell • 3GPP TS 36.201 V8.0.0 (2007-09) , TS 36.211 & TS 36.212 physical layer

description – 512 point FFT is still over-robust – Cyclic Prefix = 16.67 usec or 33.33 usec is not large enough to support large

distance communication with long echo patterns – FEC: 1/3 turbo code may have to correct for longer echoes outside the guard

interval (cyclic prefix) • Perhaps even at high received signal strength

• Received signal strength is the key driver for mobile support – Medium power medium tower SFN can support mobile devices with less

capital – However, UNICAST mobile devices need to transmit more power to reach the

towers further away

LTE cannot support sparse cell networks.

Network Support: Single large cell • 3GPP TS 36.201 V8.0.0 (2007-09) , TS 36.211 & TS 36.212 physical layer

description – 512 point FFT is still over-robust – Cyclic Prefix = 16.67 usec or 33.33 usec is not large enough to support VERY

large distance communication with long echo patterns – FEC: 1/3 turbo code WILL have to correct for longer echoes outside the guard

interval (cyclic prefix)

• Received signal strength is the key driver for mobile support – Single high power high tower may NOT support mobile devices

• Environmental factors of terrain will create ‘holes’ in coverage

– Higher transmit power will not ‘fill-in’ coverage holes inside buildings and outside severe multipath channels

• It is a physics and geometry of the market issue

LTE cannot support single large cell networks.

Broadcaster Mobile Service What Mobile Means to Us The term “mobile” as applied to ATSC 3.0 has become somewhat fuzzy and possibly overloaded. To provide greater clarity, we present the following understanding of what mobile service means to us: 1) Mobile Service expectations: We expect ATSC 3.0 transmissions to address all of the following use cases and device types: • Pedestrian handheld (outdoor) • Vehicular built-in receiver (stationary or in motion) • Fixed or portable device using external indoor antenna • Handheld or other portable (e.g., tablet) used in-vehicle • Handheld, other portable (e.g., tablet), or fixed device with embedded antenna used indoors We recognize that the range of signal levels required for successful over-the-air reception on all of the above may be wide, but our definition of mobile service would ideally include all of these use cases. Most important among these, however, is the Handheld/Portable Indoor case. 2) Maximum speed of motion: We expect that successful reception in a moving vehicle (whether using handheld/portable or built-in vehicular receivers) will be possible at ground speeds of up to 150 km/h (90 mph). 3) Mobile emergency alerting: We expect that a mobile-optimized emergency alerting modality will be included in ATSC 3.0 service. 4) Handoff: We expect that mobile service will provide a means for broadcasters to enable “handoff” of a receiver between affiliated stations or services across neighboring markets.

Definition of ‘mobile’ is different between broadcaster and wireless carriers

ATSC 3.0 Network Support: All • ATSC 3.0 physical layer description

– FFT baseline starts with 8K,16K and 32K FFT sizes and can signal other options and not overcompensate for Doppler (optimize spectral use)

– Guard Interval table starts with 12 separate entries giving a range of options to broadcasters from 28usec to over 700usec (Supporting cell radii range of 4.2-105Km).

– Scattered Pilot patterns setup for robust modes and high capacity modes for each guard interval.

– FEC starts with the most powerful known today (LDPC) to have stronger protection in difficult channels

• Higher spectral efficient capacity

– ATSC 3.0 is built from the beginning to enable all devices and architecture types (no limiting backward compatibility constraint)

• Received signal strength is the driving factor for mobile support – Indoor reception on smartphones with antenna LOSS is compensated with a range of

options: low power, medium power and high powers on different tower heights

ATSC 3.0 can support all network types and is flexible for future adaptations

ATSC 3.0 Network Optimizations

~2kWatt Small distance, close in echoes {33usec cyclic prefix optimized}

10~300kWatt

large distance, far out echoes {range of cyclic prefix options}

~2kWatt Small distance, close in echoes {33usec cyclic prefix optimized}

10~300kWatt

large distance, far out echoes {range of cyclic prefix options}

ATSC 3.0 PHY 3 Tower SFN

ATSC 3.0 PHY 6 Tower SFN

ATSC 3.0 allows flexibility for broadcasters to tune their own SFNs for best received signal strength

ATSC 3.0 Functional Architecture

Input Formatting

Coded Modulation Structure Waveform

Generation

Signaling

Return Channel

Link Layer

downlink

uplink

Control Info

Data

Data

Performance of some downlink functional blocks already shows improvement over LTE

Input and Coded Modulation Blocks

Framer Scheduler

/ Scrambler

Forward Error

Correction

Bit Interleaver Mapper Time

Interleaver

Combines multiple input streams into frame(s) with many physical layer ‘pipes’

Places frames in a selected order and scrambles data per ‘pipe’

Adds information data protection per ‘pipe’

Randomizes data bit placement within a ‘pipe’ to reduce channel’s effects

Assigns a group of data bits to a symbol per ‘pipe’

Randomizes symbols per ‘pipe’ to reduce channel’s effect (e.g. deep fades)

Signaling

Input Formatting

Coded Modulation

Input Streams

Coded Modulation Key Considerations

• Forward Error Correction Codes Function – gives receiver ability to correct errors in the transmitted signal

• LDPC (Low Density Parity Check) – state-of-the-art codes increase robustness (closer to Shannon) – Enables higher protection with less overhead

• Constellation Mapping – Range of QPSK to high order QAM enables robust and high

capacity modes – Non-uniform mapping enables more robust operation in difficult

channels (1 to 2dB gain depending on the QAM size) • Time Interleaver Function

– distributes burst noise more evenly to improve performance of FEC

LDPC FEC Performance (AWGN)

MANY operating points that are always more efficient than LTE

0

1

2

3

4

5

6

7

8

9

10

-10 -5 0 5 10 15 20 25 30

Capa

city

(bits

/sec

/Hz)

Es/No (dB)

64800 LDPC Capacity Curve in AWGN at BER=1E-6

QPSK

16QAM

64QAM

256QAM

1024QAM

A/53

LTE

Shannon Limit

7dB gain

Turbo Codes Vs LDPC – Turbo code was adopted as a FEC for 3GPP

technologies (3G, HSDPA, LTE) – LTE was designed to work with HARQ, a very low FER

is not critical (10% BLER is a typical operating point in point to point (unicast) LTE with retransmission)

– When used for Multicast/Broadcast, the absence of HARQ reduces the performance of Turbo codes.

– In ATSC3.0, there was a consensus to adopt LDPC as a FEC

– LDPC codes have a very good performance without a return channel (Quasi Error Free, suitable for broadcasting)

Structure and Waveform Blocks

OFDM Framer

Frequency Interleaver

Pilots Insertion IFFT Guard Int.

Insertion Preamble Insertion

Combines multiple inputs into single stream and format it in frames

Randomizes data cells to reduce channel’s effect (e.g. deep fades)

Inserts pilots and reserved tones for channel estimation, synchronization, PAPR reduction

Generates OFDM waveform

Inserts GI Inserts a preamble for detection, synchronization, signaling

Signaling

Structure

Waveform Generator

Waveform Key Considerations • Flexibility: network types, network sizes, service types

– Different combinations of FFT sizes, GI, scattered pilots patterns, different frame types

• Robustness – Increased signaling data robustness

• Reduced overhead to increase payload – Optimized pilots patterns for channel BW and propagation conditions

• Power savings – Time Division Multiplexing (TDM) of data frames

• Reduced complexity for easy implementation – H/W implementation and Testing

• Future extensibility – Futures extension frame parts

Physical Layer Frame Structure

• Time division multiplexing (TDM) • Each physical frame should start with a preamble • FEF allows for future technology or current other technology to be

incorporated

Preamble

Core

Preamble

Enhanced

Preamble

Preamble

Future Extendibility

…

Future Extension Frame (FEF) Frame Length

Preamble

Enhanced &

Core

OR:

Core

Preamble

Future Extendibility

…

Preamble

Enhanced &

Core

Preamble

Enhanced &

Core

Frame Length Frame Length

FEF: doorway to the future • Start with known preamble, ‘clean slate’ to put any modulation

inside that frame • Expandability for future growth • Ability to try new technology without breaking existing service and

receivers. • FEF’s enable integration of different wireless standards as various

signal formats are not restricted by PHY parameters in ATSC 3.0. – Example integration of LTE into a FEF can be seen in ‘Point-To-

MultiPoint-Overlay (P2MP) for LTE-Advanced using DVB-T’ paper by Reimers and Juretzek from IEEE-BTS conference on Oct 14, 2013.

– 802.16 Worldwide Interoperability for Microwave Access (WiMAX) example

• Can transmit FEF’s with your own technology while not breaking ATSC 3.0 service

Broadcast Overlays

10~300kWatt

ATSC 3.0: >25km, far out echoes {28…709usec of cyclic prefix options}

~40Watt WiMAX 802.16e: <4km, close in echoes {<22usec cyclic prefix}

LTE eMBMS: <10km, close in echoes {33usec cyclic prefix optimized}

~2kWatt

Preamble

Future Extendibility

Future Extension Frame (FEF)

LTE subframe

OR:

802.16e subframe

OR:

Proprietary subframe

Receiver Complexity • Desired handset feature (by broadcasters) is TV broadcast reception

– MYTH: Do LTE and it will automatically be in handset • False: Broadcaster’s frequency is different from wireless carriers’, antenna loss is worse at lower

UHF / VHF; control of tuner and other phone operation is from wireless carriers, not the user. – MYTH: adding broadcast capability to handset is trivial

• False: eMBMS from a broadcaster has above issues and additional complexity. • Note: eMBMS from wireless carriers is being tested now and search is on for content.

– Totally separate video function that will NOT impact other operations of the phone • Separate RF paths of LTE vs. ATSC 3.0

– Space is increasing on phones as displays are getting larger – PHY Demodulator chip can still be relatively small and port IP packets to an existing video decoder

already on handset • Broadcast is reception only, no transmission energy that interferes with other radios • Return link possible through IP connection via WiFi or LTE uplinks (data plans) for interactivity

– Dual Operation (receive phone calls while watching a video on the handset) is possible

• Use SEPARATE RESOURCES; either one instance of the LTE eMBMS which is hamstrung from a design for unicast data OR an instance of ATSC 3.0 PHY optimally designed (very spectrally efficient) to deliver video content.

We can make a standard to support a technical solution/option, but deployment/use in devices is based on business considerations.

Summary • ATSC 3.0 offers more flexibility and robustness than LTE.

– No network architecture is precluded, dense, sparse and single cells are enabled!

– Wide range of guard intervals and pilot patterns to choose from – Powerful FEC to protect data with different rate/robustness tradeoffs – Fully configurable to target all devices, mobile and fixed in the same

emission • ATSC 3.0 will be a flexible standard with a starting point that is more

efficient than LTE • Future Extension Frames

– Other standards can be incorporated – Broadcast overlay concepts are enabled – Experimental proprietary modulations can be tested without breaking

ATSC 3.0 service