Dec doc:IEEE b Slide 1 Submission Liang Li, WXZJ Inc. Project: IEEE P Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [ System Simulation ] Date Submitted: [ Dec.16, 2004 ] Source: [Liang Li, Liang Zhang, Yafei Tian, Chenyang Yang, Zhijian Hu, HongYu Gu ] Company: [.WXZJ Inc.] Address: [Building D, No.2, Shangdi XinXi Lu, Beijing, China ] Voice:[ ], Re: [ IEEE ] Abstract:[The analysis of orthogonal code in OPSK modulation for PHY of 915MHz and 868MHz.] Purpose:[To encourage discussion.] Notice:This document has been prepared to assist the IEEE P 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

Dec doc:IEEE b Slide 2 Submission Liang Li, WXZJ Inc. Project: IEEE P Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [ Analysis of E16 for 868/915 Band PHY] Date Submitted: [Dec ] Source: [Liang Li, Liang Zhang, Yafei Tian, Chenyang Yang, Zhijian Hu ] Company: [WXZJ] Address: [2 Xinxi St, Building D, Haidian District, Beijing, China ] Voice:[ ], Re: [ IEEE ] Abstract:[Analysis of E16 orthogonal spreading code for 868/915MHz band PHY.] Purpose:[To encourage discussion.] Notice:This document has been prepared to assist the IEEE P 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

Dec doc:IEEE b Slide 3 Submission Liang Li, WXZJ Inc. Overview This document offers key parameters of E16, and the system performances of E16 orthogonal code for 915MHz and 868MHz system: 915MHz PHY: –PSD of E16 and PHY parameters –Synchronization performance - in the presence of frequency offset –System performance Sync error, phase noise, sampling error, frequency offset, Rayleigh channel, 868MHz PHY: –PSD of E16 and PHY parameters –Synchronization performance - in the presence of frequency offset –System performance Sync error, phase noise, sampling error, frequency offset, Rayleigh channel,

Dec doc:IEEE b Slide 4 Submission Liang Li, WXZJ Inc. OPSK variants reviewed in this presentation E16 for 868MHz E16 for 915MHz Bandwidth 600KHz 1.5MHz Chip rate 400k 1M Bit rate 100kbps 250kbps Spectral efficiency Contain 99% energy in 0.79 normalized bandwidth Contain 99% energy in 1.2 normalized bandwidth Spreading16-chip seq per 4bits RF backward compatibility MSK+Tx filter or GMSK MSK CommentsQuasi constant amplitude and continuous phase Constant amplitude and continuous phase

Dec doc:IEEE b Slide 5 Submission Liang Li, WXZJ Inc. DSSS Sequence E 16 Decimal Symbol Binary SymbolChip Values The spread sequence corresponding to binary symbol “ 0000 ” is used for sync

Dec doc:IEEE b Slide 6 Submission Liang Li, WXZJ Inc. OPSK Proposal E16 Orthogonal Spreading Sequence for 915 MHz PHY

Dec doc:IEEE b Slide 7 Submission Liang Li, WXZJ Inc. 915 MHz Band PHY Key design parameters –Summary of design requirements for the TG4b PHY PSD of E16 Auto-correlation performance of E16 – Auto-correlation of O-QPSK with half sine pulse shaping / I/Q modulation at 2x sampling rate – sync in condition of frequency offset E16 Performance – simulation condition or system construction – AWGN and Rayleigh channel in ideal condition – Frame detection, synchronization, phase noise, frequency offset, sampling error, respectively (to be continued) Summary

Dec doc:IEEE b Slide 8 Submission Liang Li, WXZJ Inc. Key Parameters of E16 Bit rate 250 kBit/s –Better orthogonal characteristic –16 sequences for 4 bits mapping –Each consist of 16 chips –1M chip rate per second –Center frequency is 915MHz; Bandwidth, Pulse shape, PAPR, frequency offset –The 1 st null-null bandwidth 1.5MHz; –Half-sine pulse shape; –0dB PAPR, the same MSK scheme as 15.4, constant module and continuous phase, lower out-of- band emission; –30dB lower over 2M wide bandwidth, which satisfies the state of 15.4; –Tolerated frequency offset at least 40ppm; Multipath fading robustness –Achieve PER<10^-2 at channels with 250ns delay spread ((Multipath channel model offer by Paul with high sampling rate); Support of current RF –Support 2 MHz wide channels in the USA and other countries were they are permitted Low cost and low power consumption

Dec doc:IEEE b Slide 9 Submission Liang Li, WXZJ Inc. PSD of Tx Waveform (OQPSK+E16) Bandwidth, Pulse shape : The 1st null-null bandwidth  1.5MHz; Half-sine pulse shape: MSK modulation offers constant modulus and continuous phase; PSD 30dB lower at 1.5MHz from center frequency.

Dec doc:IEEE b Slide 10 Submission Liang Li, WXZJ Inc. PSD Characteristic PSD of OQPSK+E16 at 915Band is not affected by sampling error. Low out-of-band emission, and no need for Tx filter Satisfies the IEEE PSD requirements (in the 915 band) Source: IEEE Standard

Dec doc:IEEE b Slide 11 Submission Liang Li, WXZJ Inc. Auto-correlation performance Synchronization performance of E16 based on simulations: Auto-correlation characteristics with MSK modulation in 2x sampling rate Synchronization performance in the presence of frequency offset

Dec doc:IEEE b Slide 12 Submission Liang Li, WXZJ Inc. Auto-correlation of modulated E16 In this test, E16 spreading sequences are first OQPSK modulated with half-sine pulse shaping, and then the correlations are calculated. Auto-correlation of modulated E16

Dec doc:IEEE b Slide 13 Submission Liang Li, WXZJ Inc. Synchronization performance Simulation parameters & assumptions: –Rayleigh Channel model as suggested at TG4 discussions –O-QPSK + half-sine pulse shaping –2M sampling rate (1M chips/sec) –Frequency offset from 0ppm to 40ppm –Center frequency = 915MHz –Average over 1 million Monte-Carlo simulations Notes: 1.Synchronization is achieved by correlating local PN with received preamble impaired by frequency offset. 2.Throughout this document, the perfect synchronization (no error) in a multipath environment is defined as the receiver being synchronized to the strongest path.

Dec doc:IEEE b Slide 14 Submission Liang Li, WXZJ Inc. AWGN Model---Synchronization performance

Dec doc:IEEE b Slide 15 Submission Liang Li, WXZJ Inc. System performance Simulation parameters & assumptions: –250ns rms delay spread Rayleigh Channel model –O-QPSK modulation + half sine pulse –without frequency offset –without synchronization error –20 octets in each packet –10,000 packets for Monte-Carlo simulation –Non-coherent demodulation –No SFD detection –No fading

Dec doc:IEEE b Slide 16 Submission Liang Li, WXZJ Inc. Simulation models Discrete exponential channel model –-Sampled version of diffuse channel model offer by Paul with 4MHz sampling rate; –At least random channel realizations; –PER calculated on 20 bytes PPDUs with preamble;

Dec doc:IEEE b Slide 17 Submission Liang Li, WXZJ Inc. AWGN: Ideal Sync. vs. Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No SFD: No

Dec doc:IEEE b Slide 18 Submission Liang Li, WXZJ Inc. AWGN simulation results The BER results are close to the theoretical curve of 16-FSK. The sync error (using received signals correlated directly with local PN) has minimal effects on performance curves at low Eb/N0, and almost no effects in high SNR condition.

Dec doc:IEEE b Slide 19 Submission Liang Li, WXZJ Inc. Multiple-path Model without Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No Phase noise :No SFD: No Sync.: Correlation Down sampling error: No

Dec doc:IEEE b Slide 20 Submission Liang Li, WXZJ Inc. Multiple-path model without Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No Phase noise :No SFD: No Sync.: Correlation Down sampling error: Yes

Dec doc:IEEE b Slide 21 Submission Liang Li, WXZJ Inc. Multiple-path model without Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No Phase noise :YES SFD: No Sync.: Correlation Down sampling error: No

Dec doc:IEEE b Slide 22 Submission Liang Li, WXZJ Inc. Multiple-path model without Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No Phase noise :YES SFD: YES Sync.: Correlation Down sampling error: No

Dec doc:IEEE b Slide 23 Submission Liang Li, WXZJ Inc. Multiple-path model without Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: YES Phase noise :YES SFD: Yes Sync.: Correlation Down sampling error: No

Dec doc:IEEE b Slide 24 Submission Liang Li, WXZJ Inc. Multiple-path model without Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: YES Phase noise :YES SFD: Yes Sync.: Correlation Down sampling error: Yes

Dec doc:IEEE b Slide 25 Submission Liang Li, WXZJ Inc. Multiple-path model with Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No Phase noise :No SFD: No Sync.: Correlation Down sampling error: No

Dec doc:IEEE b Slide 26 Submission Liang Li, WXZJ Inc. Multiple-path model with Fading + Correlation Sync. Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: Yes Phase noise :Yes SFD: Yes Sync.: Correlation Down sampling error: No

Dec doc:IEEE b Slide 27 Submission Liang Li, WXZJ Inc. OQPSK Proposal E16 Orthogonal Spreading Sequence for 868 MHz PHY

Dec doc:IEEE b Slide 28 Submission Liang Li, WXZJ Inc. Key Parameters of E16 Bit rate 100 kBit/s –Better orthogonal characteristic –16 sequences for 4 bits mapping –Each consist of 16 chips –400k chip rate per second –Center frequency is 868MHz; Bandwidth, Pulse shape, PAPR, frequency offset –The 1 st null-null bandwidth 600kHz; –0dB PAPR, –GMSK modulation with r=0.2, constant module and continuous phase, lower out-of-band emission; –Nearly 50dB lower over 600kHz wide bandwidth, which satisfies the state of ETSI; –Tolerated frequency offset at least 40ppm; Multipath fading robustness –Achieve PER<10^-2 at channels with 250ns delay spread (Multipath channel model offer by Paul with high sampling rate); Support of current RF –Support current 600kHz band available at 1% duty cycle in Europe today –Allow use of extended European bands and bands in other countries once they become available Allow addition of additional 600 kHz channels as per current ETSI / ECC report (4/6 channels?) Do not expect US-like wide, unrestricted bands or all egulatorydomains –Support of more flexible channel selection method to flexibly add support for more countries Low cost and low power consumption

Dec doc:IEEE b Slide 29 Submission Liang Li, WXZJ Inc. OPSK variants reviewed in this presentation I. E16 for 868MHz II. E16 for 868MHz Bandwidth 600KHz Chip rate 400k Bit rate 100kbps Spectral efficiency Contain 99% energy in 0.79 normalized bandwidth Spreading16-chip seq per 4bits RF backward compatibility MSK+Tx filter GMSK Comments<1dB PAPRConstant amplitude and continuous phase

Dec doc:IEEE b Slide 30 Submission Liang Li, WXZJ Inc. 868 MHz Band PHY Key design parameters –Summary of design requirements for the TG4b PHY PSD of TX waveform at 868MHz RX Performance with E16 at 868MHz Band –simulation condition or system construction –AWGN and Rayleigh channel (theoretical PER results) in ideal condition –Frame detection, synchronization, phase noise, frequency offset, sampling error, respectively (to be continued) Summary

Dec doc:IEEE b Slide 31 Submission Liang Li, WXZJ Inc. Simulation models use Discrete exponential channel model –-Sampled version of diffuse channel model offer by Paul with 4x sampling rate; –At least random channel realizations; –PER calculated on 20 bytes PPDUs with preamble;

Dec doc:IEEE b Slide 32 Submission Liang Li, WXZJ Inc. Ia. OQPSK + half sine pulse without Tx filter Tx / Rx performance at 868MHz, 600KHz bandwidth E16 orthogonal code + OQPSK + half-sine pulse shaping Tx: PSD, No shaping Filter; RX: Synchronization performance Receiver (Non-Rake) performance comparison based on our simulation results

Dec doc:IEEE b Slide 33 Submission Liang Li, WXZJ Inc. 100kbps Data rate PSD 100kbps; 400k chip rate; 600k bandwidth; half sine pulse shape; No Tx filter;

Dec doc:IEEE b Slide 34 Submission Liang Li, WXZJ Inc. 100kbps Data rate performance Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No SFD: No Ideal sync

Dec doc:IEEE b Slide 35 Submission Liang Li, WXZJ Inc. Ib. OQPSK + half sine pulse with Tx filter Tx / Rx Performance at 868MHz, 600KHz bandwidth Assumption: E16 orthogonal code + OQPSK + half-sine pulse shaping Tx: PSD, 6 taps Tx digital raised cosine filter with r=0.2; Rx: Synchronization performance Receiver (Non-Rake) performance comparison based on our simulation results In the following slides, two Tx filters will be analyzed at 2x sampling rate and 4x sampling rate, respectively.

Dec doc:IEEE b Slide 36 Submission Liang Li, WXZJ Inc. Freq Response – raised cosine filter r=0.2 SUPPOSE: 1, 0.8MHz (2x)sampling rate; 2, 250kHz pass band; 3, Tx digital FIR filter; 4, 6 taps;

Dec doc:IEEE b Slide 37 Submission Liang Li, WXZJ Inc. Impulse response of Tx filter – raised cosine filter r=0.2 SUPPOSE: 1, 0.8MHz (2x)sampling rate; 2, 250kHz pass band; 3, Tx digital FIR filter; 4, 6 taps;

Dec doc:IEEE b Slide 38 Submission Liang Li, WXZJ Inc. 100kbps Data rate PSD with Tx filter 100kbps; 400k chip rate; 600k bandwidth; half sine pulse shape; 6 taps FIR Tx filter; Raised cosine filter with r=0.2; 2x over sampling rate; (0.8M sampling rate)

Dec doc:IEEE b Slide 39 Submission Liang Li, WXZJ Inc. PAPR of 100kbps with Tx filter PAPR is less than 1dB (about 0.8~0.9dB) The amplitudes of samples after Tx filter

Dec doc:IEEE b Slide 40 Submission Liang Li, WXZJ Inc. 100kbps Data rate performance Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Frame Detection: No SFD: No Ideal sync Tx filter

Dec doc:IEEE b Slide 41 Submission Liang Li, WXZJ Inc. Freq Response raised cosine filter r=0.2 SUPPOSE: 1, 1.6MHz (4x)sampling rate; 2, 250kHz pass band; 3, Tx digital FIR filter; 4, 8 taps;

Dec doc:IEEE b Slide 42 Submission Liang Li, WXZJ Inc. Impulse response of Tx filter – raised cosine filter r=0.2 SUPPOSE: 1, 0.8MHz (4x)sampling rate; 2, 250kHz pass band; 3, Tx digital FIR filter; 4, 6 taps;

Dec doc:IEEE b Slide 43 Submission Liang Li, WXZJ Inc. 100kbps Data rate PSD with Tx filter 100kbps; 400k chip rate; 600k bandwidth; half sine pulse shape; 8 taps FIR Tx filter; Raised cosine filter with r=0.2; 4x over sampling rate; (1.6M sampling rate)

Dec doc:IEEE b Slide 44 Submission Liang Li, WXZJ Inc. PAPR of 100kbps with Tx filter PAPR is less than 1dB (about 0.4~0.5dB) The amplitudes of samples after Tx filter

Dec doc:IEEE b Slide 45 Submission Liang Li, WXZJ Inc. Nonlinear PA Characteristics

Dec doc:IEEE b Slide 46 Submission Liang Li, WXZJ Inc. Impact of PA Nonlinearity: 2x sampling rate (1) Tx PSD without Tx filter or PA (2) Tx PSD with Tx filter, no PA (3) Tx PSD with Tx filter and PA Because of aliasing at relatively low sampling rate, the signal side-lobe is susceptible to PA nonlinearity. (1) (2) (3)

Dec doc:IEEE b Slide 47 Submission Liang Li, WXZJ Inc. Impact of PA Nonlinearity: 4x sampling rate (1) Tx PSD without Tx filter or PA (2) Tx PSD with Tx filter, no PA (3) Tx PSD with Tx filter and PA At 4x sampling rate, the impact of PA nonlinearity is neglectable. (1) (2) (3)

Dec doc:IEEE b Slide 48 Submission Liang Li, WXZJ Inc. II. GMSK TX/ RX Performance within 600KHz at 868MHz Assumption: E16 Orthogonal code +GMSK TX: PSD, No filter; RX: Synchronization performance Receiver (Non-Rake) performance comparison based on our simulation results

Dec doc:IEEE b Slide 49 Submission Liang Li, WXZJ Inc. GMSK sequences generated Method I : Method II :

Dec doc:IEEE b Slide 50 Submission Liang Li, WXZJ Inc. 100kbps Data rate – GMSK modulation 100kbps; 400k chip rate; 600k bandwidth; half sine pulse shape; No Tx filter; Gaussian filter before MSK modulation with r=0.2

Dec doc:IEEE b Slide 51 Submission Liang Li, WXZJ Inc. GMSK(r=0.2) 100kbps Data rate performance Packet Number: PSDU Length: 20 Byte Tx/Rx Over Sample Rate: 2 Channel Over Sample Rate: 4 Ideal sync Frame Detection: No SFD: No rms delay spread 250ns BER and PER of GMSK is calculated theoretically