1
$\begingroup$

Are there any regulations (in the US or the EU), plans to regulate or initiatives that think about the "pollution" with or amounts of laser emission in urban spaces in general and the impact it may have on health?

Given the prevalence of LiDAR in robotics and autonomous vehicles - for navigation, mapping or collision detection - I've come to worry about the safety aspect of such vast amounts of laser light that is emitted into the public space.

I know that LiDAR scanners are usually "eye safe" and infrared, but the human eye isn't the only optical sensor on earth and to other non-human participants in traffic, an array of laser sources on an autonomous vehicle (AV) or other robotic or automatic objects might "look" like a "radiation porcupine". That might not be a problem today, with relatively few AVs, but might become one in the future.

That said, I've learned in a recent discussion that LiDAR, at least with autonomous vehicles, begins to fall in popularity as the technology of choice and will probably be fully replaced in driving support systems and future production model AVs. The reason: LiDAR introduces all sorts of problems with maintainability - moving parts, transparent windows, dirt, etc. - while radar reduces mechanical complexity and gives wider design choices. The impact on health of prevalent radar is another story...

$\endgroup$
4
  • 1
    $\begingroup$ As an autonomous vehicle developer working specifically in the areas of sensor selection, I'll say that one of the big advantages of lidar is the fact that it IS a laser - the beams are pretty narrow to begin with and the beam divergence is very low. When you get a return on a lidar, you can feel pretty confident that there's actually an object at that distance. Radar devices have a much, much larger beam divergence/beam width and the tendency of radar to multipath off metallic objects can lead to false detections and makes separability very difficult. Solid state lidar is probably the future. $\endgroup$
    – Chuck
    May 22 at 0:53
  • $\begingroup$ @Chuck Interesting to hear. My last remark in my question about LiDAR making way for radar, as said, is based on discussion. I was told that a previous generation of driving assist systems of a large automobile manufacturer was based on laser, and that manufacturer moved away from it towards radar in the more up-to-date generation - due to practicability. With "Solid state lidar" you mean MEMS / micro mirror arrays, right? $\endgroup$
    – Micropolis
    May 22 at 17:15
  • 1
    $\begingroup$ I'm not aware of any commercial passenger cars using lidar for advanced driver assistance systems (ADAS). The key thing you get from radar is Doppler rangerate as part of the return signal, so you don't need to do the processing with lidar that's required - convert points to objects, perform object tracking, estimate speeds based on differential object positions, etc. You can do adaptive cruise control pretty easily with radar and lane keeping pretty easily with cameras, but once you start getting into real autonomy you need to track individual actors and that's really hard to do with radar. $\endgroup$
    – Chuck
    May 23 at 12:47
  • $\begingroup$ @Chuck Mh, I begin to question if my dialog partner did actually know that this former generation of ADAS was using LiDAR or was just (falsely) assuming it or so... +1 for your iteration on things. Thanks. $\endgroup$
    – Micropolis
    May 23 at 18:21

1 Answer 1

2
$\begingroup$

The radiation produced by LiDARs is just light in a specific frequency. One way to think of it is a very focused very small flashlight which only puts out one color, that we can't quite see. That color is also known as infrared, which is also known as heat.

The amount of light energy emitted by LIDAR sensors is quite small and there is not a significant worry about the accumulation. Picking a random outdoor rated sensor the SICK outdoorScan3 I see in the datasheet that is produces only 9.2mW on average.

LED headlights, the most efficient and comparable use in the range of 10-20W of power. (Xenon reference, paper on motorcycle headlights efficiency) So if you're accumulating 1000 LIDAR sensors on a car, the output would be the same as having 1 headlight on the car. Most cars have 2 headlights as well as a whole bunch more lights which will produce much more light radiation than the sensors.

Also for reference sunlight hitting the ground delivers over 1kW of light energy per square meter. Which would take over 100000 LIDARS pointing at the same square meter of space to produce the same effect. However most LIDAR sensors point outward and scan all around which means that the power is distributed around. Which means that at say 5m distance, only 3% go into a specific meter radius. And as such you're now looking at needing on the order of 3 million sensors around you at 5 meters away. Back calculating that to a 5 meter sphere you would need over 10000 sensors per square meter of that sphere, which is by multiple orders of magnitude higher than you could physically achieve.

And that's only to produce the level of light that the sun provides, which is a level of light that I am quite happy to walk around in.

As such I don't worry about any health effects of the many LIDARs that might be used in public spaces.

$\endgroup$
5
  • $\begingroup$ Mh, ok. The average "laser pointer" type of lasers are rated <5mW (class 1) and they are considered "eye safe", but from Internet knowledge they can burn a spot into a retina if looked into for longer than a short flash. Leaving a blind spot for life. Right? The SICK sensor outputs twice the amount of laser light, but it is rotated at high frequency thus an eye could accidentally be exposed to it only very short. That makes it safe, right? And it's infrared, which is "just heat" and doesn't burn tiny holes into retinas however focused it is? Also right? That would all be very calming. $\endgroup$
    – Micropolis
    May 13 at 1:44
  • 1
    $\begingroup$ > from Internet knowledge they can burn a spot into a retina if looked into for longer than a short flash Properly rated eye safe lasers are not generally a problem unless you stare directly into them for extended periods of time (Childrens Hospital LA). You need to watch out because you can buy imported unrated lasers pretty cheaply online. (Warnings from AAO) As a general practice it's better to just assume you don't know how strong it is and avoid it. $\endgroup$
    – Tully
    May 13 at 18:33
  • 1
    $\begingroup$ > The SICK sensor outputs twice the amount of laser light, but it is rotated at high frequency thus an eye could accidentally be exposed to it only very short. That makes it safe, right? This is part of the eye safety certification it significantly lowers the risk and they generally have things like cutoffs that disable the light output if it stops spinning. $\endgroup$
    – Tully
    May 13 at 18:34
  • 1
    $\begingroup$ > And it's infrared, which is "just heat" and doesn't burn tiny holes into retinas however focused it is? Also right? That would all be very calming Any wavelength if focused can cause damage if concentrated to a high enough power level and for long enough. Infrared usage has to be extra careful because you don't have a blink response within 1 tenth of a second. (British Journal of Opthmology) $\endgroup$
    – Tully
    May 13 at 18:41
  • $\begingroup$ Although your answer didn't point out actual law regulations, it helped in getting a generally better understanding of possible concerns, and their actual weight. The other thing seems to be that each of the used devices usually has a certification for use within public environments and as such there's no separate regulation like I was asking for. That said, only because this question hasn't yet unearthed some law doesn't mean there isn't one or should be. Thanks Tully! $\endgroup$
    – Micropolis
    May 13 at 19:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.