LIDAR-Lite: A New Benchmark for Optical Distance Measurement Sensors

This sensor technology is ideally suited for any project requiring very compact, low power, high performance distance measurement sensors, such as drones, robots, or unmanned vehicles.

PulsedLight’s novel signal processing technology enables them to offer a low-cost optical distance measurement solution with performance comparable to systems costing hundreds of dollars.

This optical distance measurement sensor is capable of measuring out to 40 meters using inexpensive, off-the-shelf, electro-optical components. The MiniModule employs the basic principles behind ToF to measure distance. However, a unique signal processing approach has been developed to determine the time delay between an encoded transmit signal and its return.

Project crowdfunding site here: LIDAR-Lite

Views: 10964

Comment by Gary McCray on January 21, 2014 at 8:36pm

This looks like a very interesting module, on their home site they actually list 4 modules some of which use Lasers.

I read about their technology previously and as I recall they are using a modulated waveform rather than strict TOF fire and measure methods.

I am very doubtful that the module shown would provide reliable performance at 40 meters under "normal" conditions, perhaps bouncing off a mirror in pitch dark though.

Objections aside, it could probably be a very acceptable scanned device for a rudimentary and very inexpensive robotic vision system and at a very attractive price.

I look forward to these actually becoming available.

Comment by Jack Crossfire on January 21, 2014 at 9:11pm

It seems to use reflected intensity + better lenses to focus the beam more than a traditional IR sensor, but it's still garbage.  It jumped off when the car's windshield passed under it.  Chris has obviously seen it. 

Comment by Gary McCray on January 21, 2014 at 10:03pm

Hi Jack,

I read about this previously and although I can't find article now, it is not intensity based like Sharp, it is actually measuring TOF but it is using a modulated wave form to do so rather than just straight fire and time return pulse method (as I recall).

But all lasers suffer from off angle reflection (just like sonar) eg: car windshield. Just light reflection versus sound reflection media.

It is probably not as robust as straight TOF, but the circuitry required is an order of magnitude less fussy too.

The Lightware guys in South Africa have the most reasonable real TOF stuff out there: http://www.lightware.co.za/

And that is mostly because they designed their own TOF IC which does all the fast stuff internally (and they actually sell the IC very reasonably).

I have one of their SF02f Laser altimeters now (today actually) which I am hoping to use for real time brushless gimbal scanned point and shoot path finding and obstacle detection.

And of course the Kinect 1 TOF version of the Kinect will be coming to Windows some time this year and while its range and sunlight performance are lacking it is a real TOF 3D point cloud camera that will really do great stuff for robots cheaply.


Developer
Comment by Randy on January 21, 2014 at 10:47pm

It's a good sensor I think.  Certainly a step up from a sonar and only slightly more expensive.  I made the video with the car actually but I think I could do better.  The surface tracking code was not customised for the sensor, I just swapped in the LED-range-finder in place of a sonar so it was only being called at 10hz and I'm sure the sensor can handle higher rates.

The arducopter code to interface with the sensor can be found here (.h file, .cpp file).

Like Gary says, it's time-of-flight base measurement and it's maximum rangle in my test was about 14m.

Comment by Gisela & Joe Noci on January 21, 2014 at 11:37pm

I use the LightWare SF02 as autolanding altimeter in our SurVoyeur UAS as standard now - it works over any terrain we have been able to fly over, sand, grass 400mm tall, gravel, bushy terrain and waterlogged salt pans. The latter if very still and smooth does give some false readings ( overrange) as the rolls in the air turbulance. The refelected laser pulse diverges and is not returned, but any small ripples and it works fine - not relevant to use as we do not land on water. What this laser does for us is completely eliminate the pressure Alt errors on the static Pressure sensor during the approach and flare phase. We used to come in on pressure alt and then at 6meters AGL we use the Senscomp ultrasound detector to detect ground and start the flare management from then. However, a mere 0.5millibar pressure error is a 4 meter height error, so that made the landing management more critical - it works fine, but requires very good pitch control, which in the thermals on hot African pans is not easy..

Using the SF02 we start the flare from 40m AGL, giving lots of height and time to get rid of the pressure alt errors.

Pressure alt errors occur due to the sensors tolerances and temp compensation ( we do full compensation on them over 0 to 85deg and have less than 0.2millbar error over that range) and due simply to changes in ambient pressure - at the coast here in Namibia ambient pressur can change 1.5millibar in a half hour..thats 12meters!

We tried having a reference pressure on ground and then sent corrections to the aircraft for the landing, and that eliminated that source of error very well, but there are other error sources least expected.

We find pressure 'pocket' variations of up to 0.8millibar over the surface of a hot pan or gravel plain. These pockets can be anything from 10meters to 100meters in diameter over the land surface, around 1 to 10 meters above the ground. This is due to the many local thermals and rising columns of hot air over the pans. That was unexpected and does influence the landings - imagine a neat approach, 8 meters above ground, and in a pocket of low pressure we suddenly think we are 6 meters higher than we should be - pitch down , only to exit the pocket and find we are now to low, etc...

The laser fixed all that! 

For the price, the SF02 is nothing short of a small miracle..

The Nampilot

PS - for those interested in ultrasound sensors - We have tried many - the maxbotics stuff, etc.They are nowhere near the performance of the Senscomp Mini-A series - they have a range of them , not sure exactly of the range they cover for each type, but we use two of them, the 0-20foot and the 0-40ft.

The 0-20ft unit we calibrate for 0-6meters, and it works! Over short grass, sand, water, very low scrub, gravel, etc. It works in 20m's wind as well, and is totally immune to vibration. It is a high voltage electrostatic pulse unit, not one of these 'puny' 10mm diameter piezo sensors. About $50 US each  the last batch was.

Comment by Phill Scott on January 22, 2014 at 2:14am
This looks interesting and I would definitely back it if it was on kickstarter (I don't like having to sign up for lots of sites that do the same thing). I've got ideas on a lightware laser for collision avoidance and am modelling the first prototype at the moment. This model would be a good surrogate for the lightware unit for the initial testing.
Comment by Phill Scott on January 22, 2014 at 2:17am
Nam pilot, thanks for sharing your experiences with these types of modules. Is the laser altimeter fixed to the fuselage or do you make sure it points vertically downwards at all times?
Comment by Gisela & Joe Noci on January 22, 2014 at 3:25am

Phill, the Laser ranger is fixed , down looking. The error on height due to A/C pitch angle is small and ignored - pitch angle during the final flare is around 6 deg @ around 1meter above ground, so error is some centimeters. I have a blog ( laser rangefinder tests in Namibia) on DIYD with pictures of the thing fitted to the plane, some descent height plots, etc..

Joe


Admin
Comment by Thomas J Coyle III on January 22, 2014 at 5:37am

Hi All,

It was my understanding that the Lightware module produces RFI that affects the GPS. Is that true?

Regards,

TCIII ArduRover2 Developer

Comment by Dan Neault on January 22, 2014 at 5:44am

Interesting idea.

Every electronic student has used phased difference to detect range, by manually adjusting receiver gain until output and input signals amplitude matches (at least they should have).

 

So through digital processing, can you adjust the gain fast enough for a reasonable sample rate, given everything that effects reflectivity?

 

Honestly not sure, but for $45 it's worth playing around with :)

 

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service