February 2004 Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Compromise for UWB Interoperability – PHY Overview] Date Submitted: [20 February, 2004] Source: [John McCorkle] Company [Motorola, Inc] Address [8133 Leesburg Pike] Voice:[703-269-3000], FAX: [703-249-3092], E-Mail:[john@xtremespectrum.com] Re: [IEEE 802.15.3a Call For Intent to Present for Ad-Hoc Meeting] Abstract: [This document provides an overview of a proposed Common Signaling Mode that would allow the inter-operation or MB-OFDM and DS-UWB devices.] Purpose: [Promote further discussion and compromise activities to advance the development of the TG3a Higher rate PHY 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. McCorkle, Motorola

Talking with each other: Basic Requirements February 2004 Talking with each other: Basic Requirements Each class of UWB devices (MB-OFDM or DS-UWB) needs a way to send messages to the other type MB-OFDM  DS-UWB DS-UWB  MB-OFDM Even better, design a common signaling mode that can be understood by either class of devices Goal: Minimize additional complexity for each type of device while enabling this extra form of communications Use existing RF components & DSP blocks to transmit message to “other-class” devices Also need to support a low-complexity receiver Lower rate mode could be acceptable if it can be used to provide robust control functions McCorkle, Motorola

February 2004 The CSM Waveform One waveform that would be straightforward for either class of device is a BPSK signal centered in the middle of the “low band” at ~ 4GHz Such a signal could be generated by both MB-OFDM and DS-UWB devices using existing RF and digital blocks MB-OFDM device contains a DAC nominally operating at 528 MHz A 528 MHz BSPK (3 dB BW) signal is likely too wide for MB-OFDM band filters Instead, DAC can be driven at slightly lower clock rate to produce a BPSK signal that will fit the MB-OFDM Tx filter Result is a 500 MHz wide BPSK signal that a DS-UWB device could receive & demodulate DS-UWB device contains a pulse generator Use this to generate a 500 MHz BPSK signal at lower chip rate This signal would fit MB-OFDM baseband Rx filter and could be demodulated by the MB-OFDM receiver McCorkle, Motorola

MB-OFDM & DS-UWB Signal Spectrum with CSM Compromise Solution February 2004 MB-OFDM & DS-UWB Signal Spectrum with CSM Compromise Solution MB-OFDM (3-band) Theoretical Spectrum Relative PSD (dB) Proposed Common Signaling Mode Band (500 MHz bandwidth) DS-UWB Low Band Pulse Shape (RRC) -3 -20 3432 3960 4488 3100 5100 Frequency (MHz) FCC Mask McCorkle, Motorola

CSM Interoperability Signal Overview February 2004 CSM Interoperability Signal Overview 500 MHz BPSK signal has similar characteristics to original pulsed-multiband signals Proposed by several companies in TG3a CFP Adopt MB-OFDM band 2 center frequency for common signaling band Centered at 3960 MHz with approximately 500 MHz bandwidth BPSK chip rate easily derived from carrier: chip rate = carrier frequency / 9 Frequency synthesis circuitry already present in MB-OFDM radio Better energy collection than wider signals (e.g. 1.5 GHz bandwidth) Fewer rake taps for equivalent energy capture  lower complexity rake Does not suffer from Rayleigh fading (>500 MHz BW) Uses different CSM piconet code for each piconet Each DEV can differentiate beacons of different piconets Provides processing gain for robust performance: signal BW is much greater than data rate Relatively long symbol intervals (55 ns) used to avoid/minimize ISI Equalization still very simple in worse multipath channels McCorkle, Motorola

MB-OFDM Transceiver Recovery of the CSM Signal February 2004 MB-OFDM Transceiver Recovery of the CSM Signal Proposed MB-OFDM transmitter architecture contains almost all required blocks for CSM signal generation Use real-valued (single) DAC clocked at 440 MHz (less than design speed) Use length-24 ternary (-1/0/1) per-piconet spreading code Result is BPSK signal with 520+ MHz bandwidth (at -10 dB points) BPSK “chip” is a “pulse” of nine cycles of a sinusoid at 3960 MHz 440 MHz DAC clock Not used for CSM Input IFFT Convolutional Bit Constellation Xmt LPF Data (9.2 Mbps w/ FEC, 18.3 Mbps un-coded) Scrambler Puncturer Insert Pilots DAC Encoder Interleaver Mapping Add CP & GI Only required if FEC is used for CSM cos ( 2 p f t ) c Apply length-24 (-1/0/1) Already present in MB-OFDM Transceiver piconet spreading code Time Frequency Code Add piconet coder (hold fixed at band 2 frequency 3960 MHz) McCorkle, Motorola

MB-OFDM Frequency Synthesis for CSM February 2004 MB-OFDM Frequency Synthesis for CSM Already present in MB-OFDM Transceiver Select DAC Clock 440 MHz Added Divider & Selector / 9 Carrier Frequency Band 2 = 3960 MHz Clock for DAC based on existing MB-OFDM PLL 440 MHz = Band #2 center frequency / 9 McCorkle, Motorola

MB-OFDM Transceiver Recovery of the CSM Signal February 2004 MB-OFDM Transceiver Recovery of the CSM Signal Proposed MB-OFDM receiver contains many required blocks for CSM reception MB-OFDM receiver contains both time-domain and frequency-domain processing Time domain processing of BPSK signal is straight-forward MB-OFDM contains correlator blocks used for synchronization functions Frequency domain processing possible using FFT engine for fast correlation Potentially allows implementation of a full symbol channel-matched filter using FFT Equalization requirements are minimal (symbol interval is 55ns, almost no ISI for CM1-3) Data processing speed is much lower due to reduced data rates (10x slower) Low-complexity BPSK demodulator can use MB-OFDM DSP blocks BPSK demodulation And FEC decoding Already present in MB-OFDM Transceiver McCorkle, Motorola

Complexity of Receiver for CSM February 2004 Complexity of Receiver for CSM MB-OFDM receiver uses I&Q sampling with 4-5 bits resolution, could be under-clocked at 440 MHz Rake receiver for CSM would have low complexity Requires relatively low data rate (18.3 MHz) 500 MHz bandwidth requires fewer taps than 1500 MHz for equivalent energy capture Requires only a few taps to collect 85% of multipath energy Only a few percent of the gate count of the MB-OFDM receiver About 8k gates at 110 MHz for 4-tap rake (at 440 Msymbols/sec) & demodulator (~2% of proposed MB-OFDM transceiver) Frequency domain processing & equalization possible McCorkle, Motorola

Simplified TX Analog Section Showing MB-OFDM CSP Signal Generation February 2004 Simplified TX Analog Section Showing MB-OFDM CSP Signal Generation Proposed DS-UWB transmit architecture contains all required blocks for CSM generation Use length-24 ternary (-1/0/+1) per-piconet spreading code Chipping rate of 440 MHz requires modification to PLL & frequency divider of DS-UWB radio Pulse forming network shown, other architectures possible Result is same CSM BPSK signal with 520+ MHz bandwidth Pulse Forming Scrambler Convolutional Encoder Puncturer Bit Interleaver Input Only required if FEC is used for CSP Apply length-24 (-1/0/1) Piconet spreading code Network 440 M Chips/sec Data (9.2 Mbps w/ FEC, 18.3 Mbps un-coded) McCorkle, Motorola

Would the CSM mode need to use Forward Error Correction? February 2004 Would the CSM mode need to use Forward Error Correction? Based on link budget analysis, an un-coded CSM mode (18 Mbps) would have less margin at 10 m than the 110 Mbps MB-OFDM But we want the CSM to be more robust, not less… FEC could be added to improve robustness, however there is no code that is common to both MB-OFDM & DS-UWB proposals MB-OFDM uses punctured codes based on a rate 1/3 k=7 code DS-UWB uses punctured codes based on a rate 1/2 k=7 code Adding FEC to the CSM could result in as much as 5 dB coding gain Would require a common code that both receivers can decode Pick one of the codes from the two proposals, or Choose a different code with relatively low complexity Following slides show link budgets for a few sample FEC choices McCorkle, Motorola

Link Budgets for CSM with Several Possible FEC Modes February 2004 Link Budgets for CSM with Several Possible FEC Modes McCorkle, Motorola

February 2004 FEC Conclusions Based on complexity versus performance trade-off analysis for convolutional and block codes to provide ~10 Mbps for CSP CSP must provide a more robust link than data modes (110+ Mbps) Requiring either MB-OFDM or DS-UWB receiver to implement additional decoder for a different convolutional code would increase complexity Further analysis is underway, no definitive recommendation at this time McCorkle, Motorola

February 2004 Conclusions The creation of a common signaling mode will allow co-existence and interoperability between DS-UWB and MB-OFDM devices Minimum useful data rate for 15.3 MAC-based interoperability is ~10 Mbps A Common Signaling Mode is described Minimal additional cost/complexity in MB-OFDM & DS-UWB Using existing MB-OFDM band 2 center frequency and bandwidth Multiple options for receive using either time or frequency domain DSP blocks in MB-OFDM radio Achieves desired data rates and robust performance Prevents coexistence problems for two different UWB PHYs Provides interoperability in a shared piconet environment McCorkle, Motorola