Electronics > Projects, Designs, and Technical Stuff
Lidar sensor accuracy versus modified laser rangefinder
(1/1)
huub8:
Hi all,
Let me first sketch the outlines of what I'm trying to do:
Please presume that I have I have a very accurate (<2cm deviation) indoor positioning system (ultra-wideband beacons).
Using this system I can locate a remote control that has a precise (<0.1 degree deviation) gyro in it. I would like to draw lines not only by moving this remote, but also by pointing it and having it track the spot its pointed at, to then visualise this in a virtual 3d space. If I know the position of the remote, the angle and the distance of the spot it's pointing at this should be possible.
I got all these things, except for a good way to accurately determine the spots distance. I was planning on using a lidar, I even got the Garmin lite v3hp lidar sensor to test it but that thing still has a 2.5cm deviation. It also has 1kHz update frequency, so way more than I need (I think). I don't really need a line of dots when moving the remote, just enough dots to be able to connect them together and get something with a reasonable resemblance to the real world. This means that I could do some averaging on the lidar to get it to be more accurate.
But still, 2.5cm devation out of the box seems a bit poor. I was wondering if there are any other lidar or in general laser range finder based solutions that allow up to preferably 10 meters (but up to 5 or so would also be usable) that have a decent update rate (20Hz might already work), and that are precise (<10 mm deviation preferably).
I did come across this guy who reprogrammed a bosh laser rangefinder.
https://www.hackster.io/iliasam/making-a-cheap-laser-rangefinder-for-arduino-4dd849
He says he even gets 60Hz stable from it for up to 6 meters with less than 10mm deviation. My question is, how do you think this compares to the Garmin v3hp in terms of reliability (the surfaces it will work on). Because I don't really care about it being bigger, or only 60Hz at < 6 meters but it does need to work reliable. Especially if I can’t do averaging.
Also, if this works so well, aren't there any more expensive laser rangefinders that could be hacked to achieve the same thing at even higher frequencies? This would seem to good to be true given the 200 euros price tag of the Garmin lidar.
Berni:
Getting high accuracy out of a laser rangefinder is quite difficult due to how fast light is. It takes it only 33ps to travel 1cm. That is a very small increment of time that is tricky to measure. Sure you can get the output to be very stable and only jitter around by a few mm but temperature and other environmental factors could affect its accuracy and introduce a offset of maybe +2cm in one extreme and -2cm in the other extreme.
Due to all this 3D tracking tends to take a different approach where you have multiple stationary beacons that measure the angle to the tracked object. These angles can then be calculated into a precise position in 3D space. Similar is concept is used in optical 3D scanning where precision is measured in thousandths of a milometer. The way this is implemented varies a lot. Movie production mocap suits have blinking LEDs on them and high speed cameras all around the room to track them. The HTC Vive uses stationary XY scanning lasers in the room and photodiodes on the tracked objects. The iPhone face scanner and the Microsoft Kinect projects a paralaxed dot pattern that is read off by a IR camera. Optical 3D scanning makes the target of interest give off light by using a projector to shine light at that part of it and then using multiple cameras to measure the angles. Etc. many ways to skin a cat.
huub8:
The thing is that the guy I mentioned and linked to seems to have done temperature compensation already.
And the speed of light part is handled mostly by the already developed electronics anyways, all I would have to do is reflash the chip and implement the same software features needed for stable measuring (which is already done by the referenced guy).
So I have no doubt that I could get a meusure rate of between 20 and 60 hertz up to 6 meters with lower then 15mm deviation, in some circumstances. But I'm wandering if these circumstance will be similar to my Garmin lidar sensor. Or if it will be way more picky.
Marco:
No idea, but this github even mentions a 280 Hz mode.
NCG:
How about 10 um? 1.50 time.
https://vimeopro.com/epicphotonics/lidar2019/video/370407331
Navigation
[0] Message Index
Go to full version