Multiple Peamble Dissagration


Just wondering what happens when multiple preambles signal reaches one DWM 1001 receiver at the same time?

Have someone met this scenario before?

I try to set up a test for this. But it seems the signal is too short. Therefore, it is quite hard to generate a such scene with couples of devices.

Thanks for the help in advance.

If devices are transmitting at the same time, then their preambles will overlap and they will interfere in the receiver… e.g. if clock offsets between the devices are small both preambles will accumulate in the receiver, this will affect the quality and accuracy of the timestamp.

Thus TDMA scheme is used in PANS SW (on-board SW that comes with DWM1001), and with it devices never transmit at the same time, each has own slot in which to transmit.

Zoran, thanks a lot. Your answer is really helpful. The TDMA scheme in PANS is awesome. I just want to check whether exists some brutal method.

Is there some advanced approaches to disaggregate the overlapped signals(Preamble)? For example, a -80dBm wins a -90dBm signal when they are overlapped.

usually the stronger signal will win, as it is larger than noise so the receiver will extract it easier, but it may not always be the case. Also if you have two devices, as their preambles accumulate, the stronger one will have a larger peak in the CIR, however the first path/timestamp may still be triggered on the smaller one if it precedes the larger one and is sufficiently larger than the noise

If we enable frame filtering, will it helps to detect the correct FP timeatamp?

Shijo Thomas

We did experiments with varying numbers of anchors deliberately sending packets at very precisely the same time, and with identical and random payloads. We even adjusted launch times to account for time of flight to the tag to make sure the packets overlapped at the tag within +/- 4 ns (the DW1000 launch time quantization).

What we found is that the DW1000 is pretty remarkable in locking onto one sender and retrieving the packet despite the overlap. This comes about mostly because the UWB modulation is a series of pulses and the receiver can find the stream of pulses that fit the packet pretty well even with a lot of other pulses happening at the same time. Signal strength was not necessarily a prime factor in which packet made it through, it seemed to be whichever packet correlated best during the preamble, or maybe whichever packet had the first pulse that arrived.

There were times and places where multiple transmitters did disrupt the ability to receive at least one of the packets, lowering packet reliability below, say, 25% (that is, the tag got at least one packet from some anchor 25% of the time). This seemed to be more position dependent than anything else as other tag positions were close to 100% reception rate.

In summary: the DW1000 tends to lock onto one sequence of pulses from one sender and get the packet received even in the face of multiple packets in the air at the same time.

These experiments were part of our research into using multiple anchors to increase range to send one message to a tag. While we can use physically large and directional antennas on anchors, tags must have small and omni antennas which limit their reception range. The goal was to get a signal to such tags without each individual anchor exceeding regulatory limits, but the combination of multiple signals would get through. Tags that operate in scheduled networks need to hear the anchors every so often to maintain network time lock.

That research is on hiatus as we do other things now, but it did result in a patent application for the idea which should issue soon.

Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA
+1 812 962 9408

thanks a lot, Mike. That is a remarkable outcome.