1. Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs)
Submission Title: [DecaWave UWB SFD Proposal]
Date Submitted: [11 February, 2010]
Source: [Michael McLaughlin] Company [DecaWave]
Address [Digital Depot, Thomas Street, Dublin 8, Ireland]
Voice:[ ], FAX: [none],
Re: [Recommendation for SFD for UWB RFID PHY]
Abstract: [Proposal describing a length 16 SFD which is very robust against bit errors in the preamble and SFD]
Purpose: [To resolve one of the TBDs in the merged IEEE f merged proposal]
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

2. Introduction
This submission discusses the criteria for a good SFD and recommends a particular sequence

3. Look in these positions
Looking for SFD
Cross correlate with SFD
Only look in pulse positions different to preamble
E.g. if SFD =
Cross correlate with
The more zeros in the SFD, the more positions you need to look at and the more chance that there will be errors.
Equivalently, the greater the number of noise sources.
Look in these positions

4. Maximise minimum distance
Need to maximise the distance between cross correlation with SFD and cross correlation with preamble.
Also must maximise the distance between and cross correlation with SFD and with transition sequence between the preamble and the SFD

5. Example Bad SFD
Small distance
Large distance

6. Minimise # occurrences of Min Distance
If minimum distance occurs only once, this is better than occurring many times
Rule of thumb:
0.2dBs worse for each doubling of # of symbols in Preamble to SFD transition sequence where min distance occurs
Translates to -0.2×Log2(# mds)
# times min distance occurs

7. SFD Scoring Formula for how good SFD is:
20Log10(Min Dist/sqrt(# zeros))-0.2Log2(#mds)
Gives performance in dBs

8. 5 places where error most likely
Time Domain proposal
Score = 3.94 dB
Equal best min dist. of 5
Ends in ‘1’
Nine ‘0’s
Noise/Error sources
Five symbols at min distance
5 error opportunities
5 places where error most likely

9. Same, but fewer error opportunities
Score = 4.22 dB
Equal best min dist. of 5
Ends in ‘1’
Two symbols at min distance
2 error opportunities
Nine ‘0’s
Noise/Error sources
2 places where error most likely

10. Best Score ending in ‘1’ Score = 4.93 dB Equal best min dist. of 5
Seven ‘0’s
Noise/Error sources
Ends in ‘1’
Seven symbols at min distance
7 error opportunities
One of eight equivalent SFDs

11. Best Score Score = 4.97 dB Equal best min dist. of 5 Seven ‘0’s
Noise/Error sources
Six symbols at min distance
6 error opportunities
Ends in ‘0’
One of twelve equivalent SFDs

12. 0101110110110001 Recommendation Best Score is 4.97dB
Implementations may not see last pulse in time to switch
Best score where last pulse is “one” i.e. does not need to be looked at is only 0.04dBs worse at 4.93dBs
DecaWave Recommended SFD is: