Two-Way Ranging Tracking Systems

1. Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Tracking systems to support Two Way Ranging.] Date Submitted: [12 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: [Using tracking information to support Two Way Ranging places demands on the receiver tracking system.] 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. TWR using Tracking information. (What is required from the tracking system.) Brethour, Time Domain

3. When doing UWB ranging, Managing crystal offsets is not optional. • Unmanaged crystal offsets lead immediately to huge errors, because the crystal drift during the long message turn-around time goes directly into the measurement. Brethour, Time Domain

4. On June 6, two different ways to manage crystal offsets were presented. 336r0: Offset management through tracking 334r0: SDS-TWR Brethour, Time Domain

5. We have two ways to address this problem. How do we choose? Only two messages on the air, BUT requires additional tracking system capability beyond just what’s required for communications. Requires more traffic on the air, BUT requires no additional capability of the receiver tracking system. Brethour, Time Domain

6. There is a natural preference for more efficient use of air time, so what’s the issue? • It comes down to a judgment of how much complexity is required in the receiver tracking loop to support the 2 message protocol. Brethour, Time Domain

7. Must an energy detect receiver have a tracking system? Homogeneous preamble signal Data = 32 octets We acquire in this part of the preamble. We are envelope aligned at the end of this time. We must stay envelope aligned for the rest of this signal time Brethour, Time Domain

8. How far might we drift?That’s a function of how long the data is on the air. • The 15.4 data section contains 32 octets • 32 octets is 256 bits of data. • At FEC rate .5 and (even with) no symbol integration, that’s 512 symbols. • 512 symbols is about .2 ms. • It won’t be less than .2 ms (no symbol integration is an outrageous assumption) Brethour, Time Domain

9. How far might we drift? • With 40 ppm crystals, the maximum relative drift rate is 80 ppm. • In .2 ms @ 80 ppm, we can drift 16 ns. • At 500 MHz, our envelope is only 5 ns wide. The useable part is only 2 ns wide. • Even a drift of 2 ns can cause us to loose our signal! Brethour, Time Domain

10. Must an energy detect receiver have a tracking system? YES! • Even if we think we will get lucky with the crystals, and only have 20 ppm of total drift, we will still drift off of a 500 MHz envelope during the shortest data payload. Brethour, Time Domain

11. So the energy detect receiver must track, how is this done? • That’s a receiver design issue, not part of our standard. • Implementers are responsible for their own receiver designs. • For discussion purposes, a suggested tracking approach is shown in the following slides. Brethour, Time Domain

12. This is an example of our signal on the air The energy detect receiver will only see this envelope. The energy detect receiver will not see this carrier. Brethour, Time Domain

13. For tracking, the envelope is oversampled For illustration, a 500 MHz (2 ns) sample rate is shown 2 3 4 5 1 The sampling system gives the tracking system, the energy contained in each 2 ns box. The energy is greatest in box 3, so the signal is said to be in box 3 Brethour, Time Domain

14. The tracking, system sees the pulse drift. 2 3 4 5 1 After some drift, the energy is greatest in box 2, so the signal is now said to be in box 2 Brethour, Time Domain

15. It is little trouble for the tracking system to count how many sample boxes the envelope drifts through over the course of a packet. • The system had to change the demodulation decision whenever the envelope energy wandered into a different box anyway! Brethour, Time Domain

16. What kind of oscillator offset correction do we expect to make? We acquire in this part of the preamble This will typically be 8 ms. We are tracking during this entire time Brethour, Time Domain

17. We are tracking the drift to 2 ns accuracy across 8 ms. That’s .25 ppm What kind of oscillator offset correction do we expect to make? We acquire in this part of the preamble This will typically be 8 ms. Brethour, Time Domain

18. .25ppm…. Is that good enough? • At least it sounds impressive! • Actually, it’s not that great. It’s a function of the energy detect receiver’s sample rate. • When we think about the error being on the order of a sample window across the entire tracked part of the packet, the error is (by definition, now) on the order of the other errors in the measurement. Brethour, Time Domain

19. What about interoperability? • Remember that in the two message protocol, both receivers measure the relative oscillator drift rate. • Either measurement can be used in the final range computation. • When a coherent receiver ranges with a non-coherent receiver, the drift rate from the coherent receiver will be used. Brethour, Time Domain

20. Review from 0336r0: At the end of the exchange, 2 measurements of the crystal offset are available A B A measures B’s oscillator drift here B embeds his measurement of A’s oscillator drift as a number in the data. (along with turn-around time) • From this slide, we will call these two measurements the Blue number and the Green number. • Radio A (whether he is coherent or energy detect) will use the number form the coherent receiver and throw away the number form the energy detect receiver. Brethour, Time Domain

21. Why was the coherent receiver’s number better? • Energy detect receiver only tracks the envelope. • Most coherent receiver designers will choose to track the carrier. Brethour, Time Domain

22. Conclusions • Even energy detect receivers will incorporate envelope tracking machinery to avoid loosing the signal during a packet. • This tracking machinery will (by it’s nature) be able to count the crossings of the envelope energy from one sample period into another. • That, plus a counter, are the essential ingredients to implement support for the two message ranging protocol. Brethour, Time Domain

23. Recommendation: • That 15.4a use tracking information to manage the crystal offsets in two way ranging. • That 15.4a keep the TWR message count to 2 messages for a ranging exchange. Brethour, Time Domain