September 2009 doc IEEE 802 15 doc 15

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Project: IEEE P 802. 15 WG for Wireless Personal Area Networks (WPANs) Submission Title: [Merged FSK Proposal TG 4 g - Update] Date Submitted: [September, 2009] Source: [Kuor-Hsin Chang 1, Rodney Hemminger 1, Bob Mason 1, John Buffington 2, Daniel Popa 2, Hartman Van. Wyk 2, Fumihide Kojima 3, Hiroshi Harada 3, Henk de Ruijter 4, Ross Sabolcik 4, Ping Xiong 4, Péter Onódy 4, Khanh Tuan Le 5, Tim Schmidl 5, Anuj Batra 5, Srinath Hosur 5, Per Roine 5, Stephen P. Pope] Company [1 Elster Electricity, 2 Itron, 3 NICT, 4 Silicon Laboratories, 5 Texas Instruments] Address: [ ] Voice: [] E-Mail: [ ] Re: [Response to CFP issued January 22 nd 2009, document 15 -09 -077 -00 -004 g ] Abstract: [This document describes the updates to the Merged FSK Proposal since the July 2009 meeting. ] Purpose: [Proposal for consideration of inclusion into 802. 15. 4 PHY draft amendment ] Notice: This document has been prepared to assist the IEEE P 802. 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 P 802. 15. Submission 1

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Merged Proposal This is a MERGED PROPOSAL from the following authors, representing a combination of equipment suppliers and Silicon vendors. This merged proposal is supported by: Submission Slide 2 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Overview • Review key philosophies of this merged proposal • Explanation of data rates • Present new technical details since the July 2009 meeting – – – Updates to band plan and channelization Simplified frame format Support for interoperability with non-standardized devices Details added for FEC correction of payload and PHY header Support for higher data rates using OFDM • Summary and conclusions Submission Slide 3 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Key Philosophies of Merged Proposal • A PHY that provides a foundation for interoperability is of utmost importance – For each frequency band, one definition – Common starting point for all communications • Based on technical merit without bias to any legacy system – Proposal is not any of the vendor’s legacy systems • Provide a robust platform with future proof data rates – Common starting point for all communications uses most robust communication scheme and data rate (i. e. lowest data rate) – From common starting point, provide mechanisms to shift to higher data rates and different modulation schemes • Core proposal based on technologies that can be implemented today, but provides a roadmap for higher data rate options without compromising interoperability – – GFSK low, medium and high data rates can be implemented today Higher data rates with OFDM can be implemented with support from silicon vendors Submission Slide 4 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Data Rate Selection Criteria • Each frequency band is optimized per regulatory requirements to provide a high data rate option based on the allowed channel bandwidth – Make each band as future proof as possible • The low data rate is the common starting point for all communications – For example, for 902 -928 and 2400 MHz bands, the low data rate is defined as: 40 kbps, GFSK, Modulation index = 1. 0, BT = 0. 5 – 40 kbps provides a robust common starting point: • • 4 d. B improvement in link budget as compared to equivalent modulation scheme at 100 kbps Can be supported by simple, battery powered devices • Legacy devices can be modified to support this data rate • Better performance with respect to multipath fading Submission Slide 5 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Advantages of a low data rate for common starting point • Multipath fading varies per environment, but 40 kbps is a good fit for the typical urban environment with a delay spread of 3 sec – To minimize ISI, the symbol rate should be less than 10% of the mean delay spread – Fs = 0. 1 / 3 sec = 33 kbaud • The urban mean spreading delay is used, but mean spreading delays are typically higher for hilly terrain, common in many rural settings • Mesh networks can typically optimize performance by choosing different communication links, BUT utility networks must also work reliably in rural areas where long communication links with limited alternate paths are required Submission Slide 6 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Importance of a Common Starting Point for All Communications • For a mesh network, one device may need to listen for messages initiated by a variety of devices. For example, a given device may need to listen for messages from: – – • • • Full function, high data rate devices with good communication links Simple devices that only support a low data rate Legacy devices that only support a low data rate Long communication hop devices that need a more robust communication link For interoperability, all devices should initiate communications using a common starting point Emphasis from NIST and others is to provide standards so that products from multiple vendors can interoperate 802. 15. 4 g by itself will not provide vendor interoperability, but it must provide a good foundation so that future standard efforts (higher layer protocols) can be added to provide the desired interoperability – Without interoperability at the PHY layer, overall interoperability will not be achieved Submission Slide 7 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Frequency Band Plans Submission Slide 8 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Updated Frequency Band Plans Frequency band 400 -430 MHz (1 MHz), 950. 9 -955. 7 MHz (Japan) 470 -510 (China) 863 -870 MHz (Europe) 902 -928 MHz (US), * Data rates and channel 2, 400 -2, 483. 5 MHz (Worldwide) Data rate Modulation Scheme Modulation Index BT Channel Spacing Data rate width vary. Scheme per frequency Modulation 50 kbps 2 -GFSK 1. 0 0. 5 200 k. Hz 40 kbps 2 -GFSK 1. 0 0. 5 400 k. Hz 40 kbps band 2 -GFSK to optimize Medium Data rate per High Data rate 100 kbps 200/400 kbps 2 -GFSK/4 GFSK 1. 0 0. 5 200 k. Hz 400 k. Hz 80 kbps 160 kbps 2 -GFSK 4 -GFSK 1. 0 1/3 0. 5 200 k. Hz 80 kbps 160/320 kbps 4 -GFSK 2 -GFSK/4 -GFSK 1/3 1. 0/(1/3) 0. 5 200/400 k. Hz 400 k. Hz 160 kbps 320 kbps 2 -GFSK 4 -GFSK 1. 0 0. 5 400 k. Hz 160 kbps 320 kbps regulatory 2 -GFSK requirements 4 -GFSK Modulation Index 1. 0 0. 5 BT 0. 5 600 k. Hz Channel Spacing Submission Low Data rate Parameter Slide 9 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> 863 -870 MHz ISM Band in Europe (1) • Frequency Band: 863 -870 MHz • Unit Channel Spacing: 200 k. Hz • Channel Spacing: N x 200 k. Hz , N=1, 2, 3, 6 • Number of channels: – 31 x 200 k. Hz – 14 x 400 k. Hz – 9 x 600 k. Hz – 4 x 1200 k. Hz • Adaptive Frequency Agility (AFA) with Listen-Before. Talk (LBT) Submission Slide 10 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> 863 -870 MHz ISM Band in Europe (2) • Modulation format: – BT=0. 5 – 2 -GFSK modulation index h=1. 0 – 4 -GFSK modulation index h=1/3 • N-GFSK Data rates: – (R 1) 40 kbps – (R 2) 80 kbps – (R 3) 160 kbps – (R 4) 320 kbps Submission Slide 11 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> 470 -510 MHz ISM Band in China (1) This proposal is presented as a possible technical solution. The suitability of this frequency band for SUN applications needs to be confirmed and aligned with the appropriate Chinese standardization bodies. • Frequency band: 470 -510 MHz • Max output power: 50 m. W (+17 d. Bm) • Channel spacing: 200 k. Hz • Transmission time: Less than 5 seconds • Frequency Hopping Spread Spectrum across the whole 40 MHz band • Dynamic power control Submission Slide 12 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> 470 -510 MHz ISM Band in China (2) # Channels Channel Spacing [k. Hz] Modulation Data Rate [kbps] Max Output Power [d. Bm] 200 2 -GFSK 40 +17 200 2 -GFSK 80 +17 200 4 -GFSK 160 +17 • Multiple sets of (offset) channels could be defined to support several co-existing networks in the same area – The main coexistence mechanism would still be the use of different hopping sequences – Although networks share the same frequency range, coexistence is improved by good far-away selectivity, as the networks have a high probability of large frequency spacing at any given moment in time – Multipath fading mitigation and coexistence with other networks are maximized utilizing the entire frequency band Submission Slide 13 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Band Plan Methodology • To accommodate potential frequency band changes and possible relaxat where Submission 14 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Simplified Frame Format Submission Slide 15 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Simplified Frame Format • • Based on input from silicon vendors, the frame structure and mechanisms to control data rate changes has been simplified Two types of PHY frames: – Type #1 – Normal frame (no data rate or modulation change) – Type #2 – Format Change frame (indicates change of data rate and/or modulation) • A Format Change frame is a simple frame followed by a Normal frame • All frames have a simple FEC algorithm to protect the PHY header Submission Slide 16 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Normal Frame Normal frame • Header FEC 1: 5 -bit wide extended Hamming code (single error correct, double error detect) covering the following 11 bits of information • Legacy: Indicates if the frame is from or to a legacy device. If Legacy = 1, the remainder of the frame is defined by the legacy vendor, but are still protected by Header FEC 1 • • Format Change: A value of zero indicates a normal frame PSDU FEC: Indicates if FEC is used for the payload. Options are provided for a simple (i. e. block) or more complex (convolutional) algorithms 0 = no FEC 1 = option #1 (RS xx, yy) 2 = option #2 (convolutional) 3 = reserved Submission Slide 17 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Normal Frame Normal frame • Data Whitening: Indicates if data whitening is used on the PSDU field. If used, the seed value is based on the channel number • • • RFU: Reserved for future use Network Id: An indication of the utility network Header FEC 2: 5 -bit wide extended Hamming code (single error correct, double error detect) covering the following 11 bit length field • Length: The length of the PSDU. The length does not include the PHY header or the CRC fields. Length is the payload size before payload FEC. Submission Slide 18 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Format Change Frame Format Change frame • Header FEC 1: 5 -bit wide extended Hamming code (single error correct, double error detect) covering the following 11 bits of information • Legacy: Indicates if the frame is from or to a legacy device. If Legacy = 1, the remainder of the frame is defined by the legacy vendor, but are still protected by Header FEC 1 • • Format Change: A value of one indicates a format change frame Setting Delay: Indicates if following normal frame (transmitted at the new data rate, modulatino scheme, etc) is transmitted after a default (0) or extended (1) settling delay. Settling delay values are functions of the Modulation/Data Rate/Channel field. • Modulation/Channel/Data Rate: Indicates the modulation, channel, and data rate to be used for the following normal frame. Submission Slide 19 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Example Frame with Data Rate Change Example Frame without Data Rate Change Submission Slide 20 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Comparison of Data Rate Options • Evaluate impact of common starting mode of 40 kbps as compared to other proposed data rates Submission Slide 21 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Comparison of Various Common Starting Point Data Rates • • Compare MAC payload size vs total frame time Analysis based on the following PHY and MAC overhead: Submission Slide 22 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Frame Time Evaluation • • The chart compares a medium data rate that starts at 40 kbps and switches to 160 kbps to a constant 100 kbps Interoperability achieved with a common starting point is not expensive! Submission Slide 23 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Frame Time Evaluation • The chart compares a high data rate that starts at 40 kbps and switches to 320 kbps to a high data rate that starts at 100 kbps and switches to 200 kbps Submission Slide 24 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> 32 -bit CRC • Propose use of CRC-32 K (Koopman) due to technical advantages • Polynomial x 32+x 30+x 29+x 28+x 26+x 20+x 19+x 17+x 16+x 15+x 11+x 10+x 7+x 6+x 4+x 2+x+1 (0 x. BA 0 DC 66 B) • Advantages – Hamming distance >= 6 up to 16 Kbit message length – Traditional CRC-32 only offers HD >=6 up to 268 bits, HD = 5 up to 2974 bits and HD = 4 up to 91607 bits. – CRC-32 K yields two additional bits of error detection for 1500 octet message lengths [1] – CRC-32 K advantage in the 269 -16360 bit (34 to 2045 octet) length range aligns with PAR requirement for minimum 1500 octet payloads [1] Koopman, P. "32 -Bit Cyclic Redundancy Codes for Internet Applications, " Int'l Conf. on Dependable Systems and Networks, 2002. Submission Slide 25 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Best data rate for common starting point Criteria Interference immunity • Must be selected for most robust option • – Best link margin – Highest immunity against interference – Best option against Multipath and ISI • Must cater for all use case scenarios Submission Interference is mitigated by – – • Time diversity Frequency diversity Improving signal to interference ratio Coding and FEC Success of mitigation is depending on the nature of interference A very simple model could be: – Uniform random density distribution of time, frequency and power • Increasing data rates – – – Improves time diversity (shorter packets) Reduce frequency diversity (larger bandwidth) Reduce signal to interference ratio (less energy per bit thus will be less probable to over shout interferers) Slide 26 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> PHY vs. MAC change management Change management Performing at • Is required for • MAC – – Data rate changes Modulation type changes Supporting legacy devices Ensuring interoperability • Needs to be done – At startup and discovery – Each time network changes (communication path change, network optimization, takeout point changes, addition/change out of units…. ) – For networks containing devices supporting multi physical layer capabilities Submission – Requires a method to multiplex between PHYs (Polling , TDM…. ) – Will reduce system capacity and increase latency (reduced system bandwidth to 25% with TDM for 2 PHY implementation) • PHY – Offers ultimate flexibility for management – Have very little overhead – Proposal outperforms single PHY implementation from CPP Slide 27 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Support for Legacy Devices 1. Propose a method for supporting any legacy device ü existing and ongoing deployments will not become obsolete ü simultaneous (and parallel) operation of any system based on legacy and standard devices, respectively 2. Propose a method that opens up for multi-vendor interoperability 3. Minimize the impact of legacy device support on the standard and not encumber the choice of the “best” technology Submission <Elster>, <Itron>, <NICT>, <Silicon Labs>, Slide 28 <Elster>, <Itron>, <NICT>, <Silicon <Texas Instruments>, <Stephen P. Pope>

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Support for legacy devices (cont’d) q Upgrade over-the-air the legacy devices affected by 802. 15. 4 g support ü only legacy devices that can accommodate radio parameter changes, while keeping (transmission link) communication performance at an acceptable level q Let system implementations decide if standard devices support legacy devices ü standard devices can support legacy devices by dual-stacking (proprietary layers and 802. 15. 4(e)g layers) rather than bridging q BUT ü Make standard PHY able to recognize if legacy devices are present in the field by using standard information for legacy device identification (i. e. , format frame change) ü modulated with the common starting point : 2 -GFSK, 40 kbps ü respects all PHY parameters as defined in this proposal ü Give PHY Layer tools to support a cross-layer efficient interoperability Submission <Elster>, <Itron>, <NICT>, <Silicon Labs>, Slide 29 <Elster>, <Itron>, <NICT>, <Silicon <Texas Instruments>, <Stephen P. Pope>

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Support for legacy devices (cont’d) Legacy Device (LD) Upper Layers upgraded (legacy) MAC** (upgraded) PHY** 802. 15. 4(e)g standard device (SD) with dual-stack Upper Layers upgraded legacy MAC 802. 15. 4(e)g SD 802. 15. 4 PHY* Upper Layers 802. 14. 5 MAC 802. 15. 4 PHY (*) From the perspective of 802. 15. 4 g, standardize only the transmission of some PHY fields required for legacy device identification purpose; however, vendors can accommodate multiple PHYs (other than 15. 4 g) on such devices, based on their legacy system parameters; for flexibility, this should be a vendor prerogative (**) over-the-air upgrade the legacy MAC that will further (re-)configure legacy PHY to deal with radio parameter changes for supporting legacy device identification Submission <Elster>, <Itron>, <NICT>, <Silicon Labs>, Slide 30 <Elster>, <Itron>, <NICT>, <Silicon <Texas Instruments>, <Stephen P. Pope>

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Support for legacy devices (cont’d) PPDU format supporting legacy devices, modulation, data rate, PHY parameters, … User case A: “Shim” legacy PHY upgrade Format Change Frame Settling delay • most common modulation: 2 -GFSK • lowest acceptable and robust data rate : 40 Kbps • respect all (PHY+MAC) parameters as defined by 802. 15. 4(e)g, e. g. , channel spacing, channel bandwidth User case B: Full legacy PHY upgrade Submission Format Change Frame Legacy frame format • data sent with respect to some specific (legacy) PHY parameters • legacy PHY (and its parameters) to be defined by each vendor but not standardized Legacy frame format <Elster>, <Itron>, <NICT>, <Silicon Labs>, Slide 31 <Elster>, <Itron>, <NICT>, <Silicon <Texas Instruments>, <Stephen P. Pope>

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Support legacy devices with 802. 15. 4 g PHY 802. 15. 4 g devices receiving frames from legacy device Format Change Frame processing @ “std” PHY parameters No Legacy device 2 No “Discard” frame 1 Process frame following Format Change Frame as defined by each vendor YES Support legacy devices ? YES Idle 2 Frame processing @ standard PHY parameters Idle 1 Idle Submission <Elster>, <Itron>, <NICT>, <Silicon Labs>, Slide 32 <Elster>, <Itron>, <NICT>, <Silicon <Texas Instruments>, <Stephen P. Pope>

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Advantages 1. Minimum impact on standard development ü minimum on-air cost, minimum complexity and can be ignored where not necessary 2. Does not require “bridging everywhere” to support legacy devices ü where possible just over-the-air upgrade the legacy devices 3. Opens up for multi-vendor interoperability ü open platform by stacking up multi-vendor protocols on top of a common PHY (and MAC) 4. Provides extensibility ü further versions of the 802. 15. 4 g PHY standard (different modulation) can be supported Submission <Elster>, <Itron>, <NICT>, <Silicon Labs>, Slide 33 <Elster>, <Itron>, <NICT>, <Silicon <Texas Instruments>, <Stephen P. Pope>

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Payload FEC Design for Merged FSK Proposal Stephen P. Pope spp@rahul. net Submission Slide 34 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Forward Error Correction Algorithms • Standard PHY header requires simple FEC • Optional FEC for PHY payload – Will present Reed Solomon option for PHY payload FEC Submission Slide 35 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Payload FEC Design -- Goals • Provide significant normalized coding gain in AWGN • Low gate-count / low-power • Reasonable latency (to meet turnaround time) • Not too esoteric -- well known coding preferred • Systematic code more desirable • Must allow for payload sizes from 1 -2048 octets • Burst-error capability not a specific requirement (but may add to robustness) • Variable code rate not a requirement (selectable data rates already available in modem) Submission Slide 36 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Overview of Code Design Tradeoffs • Some popular codes are too complex for this application: – Near-channel capacity codes (e. . LDPC, Turbo) – Concatenated codes (e. g RS over convolutional) • Convolutional codes are workable – Good performance at low SNR but soft-decisions needed – Performance less interesting at higher SNR or for longer payloads – High gate-count • Algebraic codes – Many low gate-count possibilities: BCH, Golay, Reed-Solomon – Reed-Solomon codes meet all requirements and can be low gate-count Submission Slide 37 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Properties of Reed-Solomon Codes • Many choices of RS parameters m, n, k m = symbol size n = total symbols per codeword k = information symbols per codeword t = (n-k)/2 = correction ability n < 2 m for non-extended codes n = 2 m extended code n = 2 m + 1 doubly-extended code n > 2 m + 1 Algebraic Geometry Code these are increasingly esoteric • Tend to perform best when payload size (in bits) is roughly the same order as the maximum codeword size n * m For our range of packet sizes this suggests m = 6 to m = 8 • Approach: evaluate RS code parameter possibilities and find an economical, effective set of parameters Submission Slide 38 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 Submission doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Slide 39 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 Submission doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Slide 40 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 Submission doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Slide 41 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Performance Comparison: PER vs. Eb/No for 6 - and 8 -bit RS Codes Assumptions: BER vs. SNR derived for 2 -FSK modulation AWGN Channel Errors-only decoding (no erasures) Submission Slide 42 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 Submission doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Slide 43 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 Submission doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Slide 44 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 Submission doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Slide 45 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Summary of Performance Comparisons (6 -bit vs. 8 -bit RS in AWGN) Coding Gain – payload length 28 The 6 -bit, RS(52, 38) code has a normalized coding gain of 3. 4 d. B, compared to 2. 9 d. B for the 8 -bit RS(38, 28) code. Coding Gain – payload length 180 A long, 8 -bit RS(240, 180) code has 4. 7 d. B of normalized coding gain, compared to 4. 1 d. B for the 6 bit code, or 3. 5 d. B for the short, 8 -bit (38, 28) code. Coding Gain – payload length 1500 The 6 -bit, RS(52, 38) code has a normalized coding gain of 4. 7 d. B, whereas various 8 -bit RS codes range from 4. 9 to 5. 3 d. B normalized coding gain. Submission Slide 46 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Support for Higher Data Rates Using OFDM Tim Schmidl – Texas Instruments Submission Slide 47 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Support for Higher Data Rates Using OFDM • OFDM can be deployed in a network to increase data rates – Can be a secondary deployment when the installed base is FSK • Channelization for OFDM should be compatible with FSK – Should be the same channel bandwidth as FSK or integer multiple of the bandwidth • Higher data rates provided by OFDM • How to transmit OFDM when FSK is frequency hopping Submission Slide 48 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Channelization for OFDM and FSK OFDM bandwidth FSK Low data rate FSK Medium data rate FSK High data rate 281. 25 kbps 200 k. Hz (16 pt FFT) 400 k. Hz (32 pt FFT) 600 k. Hz (64 pt FFT) 281. 25 kbps 562. 5 kbps 750 kbps 200 k. Hz 400 k. Hz 600 k. Hz 902 -928 MHz (US) 200 k. Hz (16 pt FFT) 400 k. Hz (32 pt FFT) 800 k. Hz (64 pt FFT) 281. 25 kbps 562. 5 kbps 750 kbps 200/400 k. Hz 400 k. Hz 950. 9 -955. 7 MHz (Japan) 200 k. Hz (16 pt FFT) 600 k. Hz (64 pt FFT) 281. 25 kbps 750 kbps 200 k. Hz 2, 400 -2, 483. 5 MHz (Worldwide) 200 k. Hz (16 pt FFT) 400 k. Hz (32 pt FFT) 800 k. Hz (64 pt FFT) 281. 25 kbps 562. 5 kbps 750 kbps 200/400 k. Hz 1200 k. Hz (128 pt FFT) 750 kbps Frequency band 426. 025 -469. 4875 MHz 200 k. Hz (16 pt FFT) (Japan) 863 -870 MHz (Europe) TV White Space Submission OFDM max data rate Slide 49 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Single OFDM Option for Each Band • There are 5 OFDM options defined with FFT sizes of 128, 64, 32, 16, 8 • 5 different preambles, one for each option. Each preamble is easily generated using the IFFT. • It is inefficient to search for multiple preambles – Assumed that when the network is deployed there will be one channelization defined for OFDM for a particular band so that the option used is known is advance Submission Slide 50 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> PHY Header for OFDM • The PHY header does not have to be the same for FSK and OFDM • There are 36 bits in the PHY header • 4 reserved bits to allow for future evolution of the standard • 5 RATE bits indicate which of the 7 MCS levels is used and allows for future expansion • LENGTH is 11 bits • SCRAMBLER seed is 2 bits since usually no more than 4 retransmissions are needed for a packet. After 4 transmissions the seed can be reused • PHY header is protected by the same convolutional code as the data Submission Slide 51 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Questions? Submission Slide 52 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Backup Slides Submission Slide 53 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 Submission doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Slide 54 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Optional Data Whitening • • Enabled by default; can be disabled with extended PHY header Seed value based on channel number – no PHR overhead required 8 -bit additive scrambler, using LFSR with feedback polynomial x 8 + x 6 + x 5 + x 4 + 1 Yields maximum length sequence (28 - 1) Submission Slide 55 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Link budget Submission Slide 56 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Multipath Fading and Data Rate • Fixed (non-time variant) fading characterized by mean delay spread (average value of the delayed signal components) • Mean delay spread will vary based on environment, typical values are as follows: Source: Submission Lee, William C. Y. , “Mobile Communications Design Fundamentals”, John Wiley & Sons, Inc, 1993 Slide 57 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Multipath Fading and Data Rate • For minimal ISI, the symbol rate (Rb) should be less than 10% of the mean delay spread. In other words: Rb < 0. 1 * Δ • 1 Using the above rule, and the mean delay spreads for various environments, the recommended baud rates for each environment can be calculated: 1 Source: Submission Lee, William C. Y. , “Mobile Communications Design Fundamentals”, John Wiley & Sons, Inc, 1993, p 38 -41. Slide 58 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Multipath Fading and Data Rate • The variability in mean delay spreads based on environment presents a strong argument for: – 3 data rates to minimize multipath fading effects – 4 level modulation to achieve higher data rates without reducing system throughput and reliability Submission Slide 59 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Power Spectral Density 2 -GFSK, BT = 0. 5, 160 kbps Trace 4 (Green) • Mod Index = 0. 75 • 20 d. B BW = 192 k. Hz Submission Slide 60 <Elster>, <Itron>, <NICT>, <Silicon

September 2009 doc. : IEEE 802. 15 -<doc 15 -09 -0628 -01 -004 g> Spectral Density 4 -GFSK, BT = 0. 5, 320 kbps • Freq Separation = 80 k. Hz (-120, -40, +120) • 20 d. B BW = 373 k. Hz Submission Slide 61 <Elster>, <Itron>, <NICT>, <Silicon