Hacker News Reader

Comment by scotty79 | Hacker News Reader

I wonder how well this works if 20 such devices shine on the same object.

Lidars can interfere but somewhat less than other types of active sensors. First of all, the detector only needs to be on for a microsecond or two, and it's unlikely for two sensors to be scanning at the same microsecond. Second, laser spots are fairly small and unlikely to overlap. Finally, there are techniques to further get rid of crosstalk, such as coded random pulses with a matched filter.

> First of all, the detector only needs to be on for a microsecond or two, and it's unlikely for two sensors to be scanning at the same microsecond.

This is subject to the Birthday Paradox, right?

LIDARs for volume use should add a few microseconds of random (not pseudorandom) jitter to the outgoing pulse time. That will prevent multiple interfering scans from different units from all synchronizing. You may still get blinded on one scan, but not all of them.

The duty cycles of lidars are generally low enough that interference is rare. The big problem in my experience is objects glinting in sunlight which isn't a problem if you're in in a warehouse but is for rooms with big windows or outside. In any event you always need to be filtering your inputs rather than taking every reading as a sure thing.

Honestly the sunlight issue has been variable for me. Velodyne processing seems to deal with that better than robosense.

That's sort of like asking how well cameras would work if 20 of them were looking at the same object.

Lidar sensors can interfere with each other, and a certain industry-standard company is famous for having terrible problems with this issue, but there are engineering solutions to this problem. FMCW is a popular choice, and gives the benefit of providing instantaneous velocity readings. Of course, due to the Heisenberg uncertainty principle, this means you also get worse distance estimation. There are other ways to engineer around the interference problem as well.

https://www.laserfocusworld.com/home/article/16556322/lasers...

> like asking how well cameras would work if 20 of them were looking at the same object.

> Lidar sensors can interfere with each other

Then it's not really like 20 cameras looking at the same object. Mostly due to the fact that cameras are passive observers and not really emitting much. And if they do, it's light and it won't really affect the others because they all benefit from it (within reason).

I'm reasonably certain that such issues will still appear in productive use in the future but will get fixed at the time, not in labs today.

Fair enough, I didn't really expand my comment enough to make my point and was about to delete it but now that someone replied I'll just leave it as is since I don't want to write a small essay.

A good lidar sensor won't have issues with interference.

Is it more or less a forgone conclusion that a good LiDAR sensor would not be affected by potentially hundreds of other similar or identical ones shining their beams across the same region?

For FMCW, there should be no interference at all for any number of sensors (there is more information in the link in my original comment). For pulsed lidar, it's an issue of how well you engineer it. I can confidently say a good sensor can regularly withstand dozens of sensors operating simultaneously in a small area because I've personally witnessed it. The devil is largely in the details of specifications and manufacturing quality (that latter one is a much bigger issue in reality), but there is no theoretical reason you couldn't make it work with hundreds of sensors simultaneously. And like I said, I know current off-the-shelf parts that will work with dozens of sensors simultaneously.

Maybe I can add a little color to my original comment this way: most lidar sensors today, including some very expensive ones from supposedly reputable vendors, are not very good. In my experience, it is more often a manufacturing problem than a design problem (this varies more by company than it does by technology).

Imagine every car has a FMCW LIDAR. The chance of interference is very high. Additionally if those receivers are using SE PDs, then these sensors will also be “blinded” (TIAs and or PDs will saturate) when a nearby Lidar system shines its laser into the detector.

I suppose it will be a good while before we can approach an empirical proof for this sort of thing, since FMCW lidars are still very scarce, even more so than pulsed lidars right now. However, even if every car does have an FMCW lidar, the conditions required to get them to interfere with each other is are:

a) Have identical laser wavelength. Not just '905nm' or '1550nm', but _precisely_ the same wavelength. This is very hard to do even if you try.

b) Have a coincident beam path. Again, this needs to be very precisely aligned.

c) Have an overlapping coherence area. This is a bit technical, but it is a higher bar than just having spots spatially overlapping.

d) Have coherent+matching phase fronts at the detector. Again this is a fairly technical subject these properties vary along the beam path, and transversely. This also vary in time, temperature and many other things. The source lidar is able to 'interfere' with itself (in other words, get a signal), because it compensates for all of these effects with a local copy of the outgoing laser light. Other lidars' outgoing beams will in general, even for 100 cars, not be 'synced up' in this way.

Moreover, those conditions are just the intrinsic interference rejection properties of coherent lidars. Layered on top of that is that two lidars need to be using the same type of modulation, bullseye each other as they scan around the FOV, and provide enough photons to actually contribute to the signal. Then, if you satisfy all of those prerequisites, the interfering lidar also needs to overcome any heuristic/algorithmic rejection of spurious signals. Finally, if all of those conditions match up and you get a signal to punch through, and it's strong enough to over come the true signal, and you can't tell that it's an erroneous signal, then it will result on one bad/missing point in a frame of thousands of points, present for one frame.

You're correct, however, that there is a saturation issue. If you just DOS the photo diodes with photons you can potentially prevent any signals from getting through. But again, this isn't super easy to do. The detectors will almost certainly be balanced, not single ended, and AC coupled. So you really have to blast the photo diode, effectively bringing it up to it's damage threshold so it is just flooded with current and can't do anything, and/or just breaks. The raw laser light doesn't do much, both because the DC signal is rejected and because the balanced detectors will reject common mode signals (clearly you know this already). You also have the same issue with needing to shine into a very narrow field of view, at the right time, for long enough to matter.

Unfortunately FMCW lidar is sweeping the lasers across the same band. You don’t need the exact frequency just a beat frequency that’s within your detection Bw.

Also balanced detectors have something called common mode rejection. This is not infinite. In high volume applications it’s difficult for this to be >25dB but you can buy some devices >35dB.

Given that Lidar dynamic range is ~100dB you will definitely see the DC. I’ve not thought about this too much but it seems like an issue for the AGC as your demodulator won’t be bothered by it.

It's true that the laser frequency is sweeping, but it very well may not be over the same band. The sweep bandwidth in a typical lidar is likely in the 1-10GHz range. The carrier frequency of the laser that this modulation is riding on is probably in the neighborhood of 200THz. Let's say you're using a telecom laser at 1550nm. The actual wavelength of that laser will centered on some channel in the 1530-1580nm band, with each channel spaced by say 100GHz. So already each laser might intentionally be in a different channel, depending on chance and how many cars are there. But even if they are in the same channel, the chirp bandwidth is small compared to the channel bandwidth, so there will likely be at most only partial overlap, depending on where the respective center frequencies actually are. Unless your lidar is using a very expensive, very fiddly laser system, this center frequency will be drifting around within the channel all the time. It varies with temperature, mechanical stress, output power and a bunch of other stuff, depending on the type of laser. However, even if the lasers are magically in the same channel, and perfectly locked to the same center frequency, you still need the light be coherent to produce an interfering RF signal. They will not be coherent.

Certainly the balanced detectors will have finite CMRR. In general you definitely have to make a good detector but it doesn't need to reject to 100dBc. A photodiode might have 100dB of dynamic range, but most likely your RF front end does not, and more importantly for most applications you will be dominated by photon shot noise, so you don't need to push common mode signals all the way to your electronic noise floor. 35dB of rejection works wonders.

I'd imagine that those cameras might have trouble if they were all using long exposures and strobe lights.

As someone outside of the industry, what's the company that has problems with interference?

When you get dozens of photographers shooting the same subject and all with their own flash, yes you can get interference that wrecks certain images. And the more photographers gather together the more of them will interfere. So indeed your comparison is apt and doesn't help me understand why it isn't a problem. The non-flash scenario is of course not applicable since lidar is actively lighting its subjects.

> Of course, due to the Heisenberg uncertainty principle, this means you also get worse distance estimation.

This isn't how the Heisenberg uncertainty principle works. For macroscopic objects, the effects are completely dwarfed by other phenomenon. Keep in mind that Planck's constant is 10^{-33} meters.

I was actually wondering about this the other day. So what equation(s) would you use to determine the variance in distance estimation relative to velocity estimation? And it sounds like you're strongly implying the distance variation is immeasurably small while accurately estimating velocity - is this correct? I'm not sure the macro point makes sense, since you could have a large object with only one point measuring it (or more realistically a dozen points, but still far from what people mean when they say macro). But I'm curious to learn more if you can provide the math.

I'm pretty sure the effect you are discussing has to do with the uncertainty relationship inherent to the Fourier Transform [0]. This is very closely related to the Heisenberg uncertainty principle, and states you cannot simultaneously constrain time and frequency, which are the values you need to measure for position and velocity, respectively. In the context of signal processing applications, I don't think the particle nature of light is typically considered, which is why it may not be exactly correct to refer to it as the Heisenberg uncertainty principle in this context. This is a bit outside my domain though, so take it with a grain of salt.

[0] https://en.wikipedia.org/wiki/Uncertainty_principle#Signal_p...

So your're correct that there is a Fourier Transform analogy for the uncertainty principle, but in the context of FMCW lidars (which brought up the question of velocity vs position uncertainty), the measurement of frequency actually determines both the position and the velocity. It's actually a problem for most FMCW lidars because you just get 1-2 frequency measurements and somehow need to disentangle what the range frequency is, as well as what the doppler (velocity) frequency is. A massive amount of effort has been put into developing lidar methods and architectures that solve this problem well.

But in summary, the uncertainty principle as encountered in quantum mechanics has ~nothing to do with a trade off between range accuracy and range uncertainty. It's possible that it could come into play in a very detailed treatment of FMCW lidar SNR, in the context of counting return photons, but also not generally necessary there. The time-frequency uncertainty plays a role in that the range and velocity resolution both get better the longer you stare at a signal. So for a given amount of reflected light, at a given range/velocity, there is a fundamental lower bound to how long you must integrate to a) get a signal at all and b) achieve a desired precision.

It's not just an analogy--the underlying math is the same. These course notes have a nice little summary + a proof: http://www.its.caltech.edu/~matilde/GaborLocalization.pdf

Thank you for this! This is exactly what I was looking for.

It seems to be an extract from "Foundations of Field Computation" by Bruce MacLennan, if you want to read the whole thing: http://web.eecs.utk.edu/~bmaclenn/FFC.pdf

He and Dr. Marcolli have a bunch of interesting stuff on their websites if you like this sort of stuff.

The position/momentum pair is just one (important) case of a more generalized Uncertainty Principle. There's a similar one, due to Gabor (1946), which says that you can't have perfect time-limited and band-limited information, which is presumably what the OP is referring to.

The underlying math is the same, and there's a principle of complementarity that describes other pairs of quantities that need to be traded against one another.

Reply on news.ycombinator.com