Ranging with Energy-Detect Receivers

1. Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Ranging with Energy-Detect Receivers] Date Submitted: [26 June, 2005] Source: [Vern Brethour] Company [Time Domain Corp.] Address [7057 Old Madison Pike; Suite 250; Huntsville, Alabama 35806; USA] Voice:[(256) 428-6331], FAX: [(256) 922-0387], E-Mail: [vern.brethour@timedomain.com] Re: [802.15.4a.] Abstract: [Comments on 15-05-0336-00-004a.] Purpose: [To promote discussion in 802.15.4a.] 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. Brethour, Time Domain

2. Ranging with Energy Detect Receivers. Comments on 15-05-0363-00-004a Brethour, Time Domain

3. 05-0363r0 shows the energy detect ranging performance hitting an error floor at 17 dB Brethour, Time Domain

4. 17 dB is a lot of S/N in a receiver that can’t do coherent integration. • So how is the 17 dB going to be achieved if coherent integration is not available? • Simple! By reducing the range. Brethour, Time Domain

5. So what does the range come to if we need 17 dB of S/N at the receive end? • It depends on the channel. • There is no simple single answer. • But rather than say NOTHING about the range, I will offer a simple single answer! Brethour, Time Domain

6. The single simple estimate comes from a link budget type analysis. • 15-05-0371-00-004a is a companion spreadsheet to this presentation. • I will argue that for a general ballpark type estimate of performance, this link budget analysis is actually better than any one simulation. • It is not better than 800 simulations of channels which are appropriately calibrated for the lengths we are talking about. Brethour, Time Domain

7. 05-0371r0 looks like this: The path loss exponent goes in this cell The “needed” 17 db goes in this cell When our link margin crosses zero we are at the maximum link distance. Brethour, Time Domain

8. So what does it say? • If we need 17 dB of S/N to do our ranging, we’ve got a 2 meter link (assuming a path loss exponent of 2 for the first meter and 3.5 for the second meter). • That’s not right: Because, with a 2 meter link, we won’t have a path loss exponent of 3.5! Brethour, Time Domain

9. So what if the path loss exponent is 2? • With a path loss exponent of 2, we predict a 3 meter link. • Path loss exponent of 2 is as good as it reasonably gets. Brethour, Time Domain

10. So we have a 3 meter link, is that Okay? • We always said that the energy detect receiver would operate at reduced range from the coherent receiver. • We shouldn’t be surprised, 17 dB is a lot of S/N. Brethour, Time Domain

11. What kind of range errors are we getting on this 3 meter link? The error floor is showing up at 2.5 ns. Brethour, Time Domain

12. So we have a 3 meter link, is that Okay? • We need to wonder! • We’re looking at range errors of almost a meter on a 3 meter link. • With node densities that high, we almost don’t need to talk about GDOP, but if we did use a GDOP factor of 4, we’d be looking at 2 meter errors on a 3 meter link. Brethour, Time Domain

13. Okay, so the performance isn’t great.. But at least it’s a simple receiver….. Right? • No it’s not! • To do the leading edge detection requires (at least) two receive channels with a controllable known time offset between them. Brethour, Time Domain

14. 2 channels…. What’s that about? • One channel tracks the acquire point & the other does leading edge detection. Receive chain 1 Acquire and track LNA Controllable offset Receive chain 2 Leading edge detection Brethour, Time Domain

15. 2 channels…. Is that really necessary? • No. That’s the analogue approach. With the analogue approach, samples are kept track of at the symbol rate. • FT has talked about continuous sampling and keeping track of all the samples. That’s a digital approach with only one receive chain, but now samples must be kept track of at the tracking (envelope over-sample) granularity. Brethour, Time Domain

16. So why am I saying it’s not a “simple receiver”? • Dangerous assertion: “Simple” is in the eye of the beholder. • If a “simple receiver” is defined to be the simplest receiver that can do communications, then by that metric, the ranging capable receiver is not simple. Brethour, Time Domain

17. At least we have a long (nominal 4 ms) preamble available to try to get some processing gain. (?) • Not really. • The 4 ms must be used to search the entire leading edge search back zone. • The good news is that if we are only putting in 3 meter links, the search back zone is relatively small. Brethour, Time Domain

18. Covering the search-back zone. • Here I’m assuming that the receiver is using the analogue approach. • For a 3 meter link, let’s assume a 24 ns search-back zone. • Even with a 3 ns search step size, that’s 8 search offsets that must be characterized. Brethour, Time Domain

19. Characterizing the search-back steps • If there are 8 steps to be characterized, then the preamble is broken up into 7 segments to be covered by the search channel. Assuming that .5 ms of the 4 ms preamble was spent on acquisition and establishing tracking lock, then only 3.5 ms remain to be split 7 ways. • Only half a ms is available for each characterization. Brethour, Time Domain

20. Conclusion. • The expected link distances for ranging with energy detect receivers are on the order of 3 meters. • The expected range errors with energy detect receivers are on the order of a meter. • If the communication only receiver is our metric for a “simple receiver” then the receiver necessary to get this ranging performance is not a simple receiver. Brethour, Time Domain