#203 Best LIDAR Sensors for Makers (Comparison and Test)

first measurement of distances became a critical technology because it enables together with artificial intelligence autonomous movements of cars or planes today we will get an overview and compare readily available devices for makers gritty youtubers here is the guy with the Swiss accent with a new episode and fresh ideas around sensors and microcontrollers we all know that Google and other companies use laser light our systems on the roof of their cars these are multi thousand dollar machines and not reachable for an ordinary maker more than a year ago in video number 119 I got a tiny chip which measured the distances up to one meter using laser light it introduced me to the term time-of-flight a miracle in my opinion however I'm too fast let's first get an overview the simplest way of measuring distances with light is to measure the intensity of the reflected beam such sensors output an analog signal and you can find the distance with this curve they are not very accurate and can be disturbed by any source of infrared light if you mount one of those on a servo you get a reasonable obstacle avoidance device for under ten dollars you measure distances already the old Greeks used triangulation and all sailors used it to find their positions for our purpose we send a light beam towards an object the angle of the reflected beam depends on the distance between the object and our sensor if we project this beam on a one dimensional camera we can calculate the range according to the place the camera sees the reflected beam I have no such sensor in my lab and therefore will not cover it in this video a second technology is also known for a while measuring phases of waves each frequency has a defined wavelength and you can determine a distance based on the phase shift of a reflected signal the last method is to measure the time a wave needs to travel a distance we know there's principal since our childhood when we saw lightning we started to count 21 22 23 and stopped when we heard the Thunder we multiplied the seconds with three hundred meters and got our distance to the lightning why did this work because sound travels around 300 meters per second in air light moves much faster 300,000 kilometers per second which was at that days indefinitely fast so we saw the Lightning immediately and heard the Thunder with the delay of the distance we use a similar principle with our ultrasonic sensors just without thunder a loudspeaker sends a tone and the device starts to count as soon as the reflected sound hits our microphone the card stops again we can multiply this time by 300 meters and get the distance simple some crazy engineers had the idea to do the same thing with light the advantage would be a minimum angle of measurement of a few degrees and if we use laser light we can measure long distances with quite low power and your doc is not disturbed by high frequency sound so let's summarize the different possibilities for unprecedent we can measure the intensity of the reflected light below 10 meters distance we can use triangulation with light it is cheap and accurate in the range till 100 meters we can measure phase shift it has a low manufacturing cost but it is slow above 100 meters and if we need fast measurements we have to measure the time of flight of laser light this is extremely fast and until now the most expensive method let's look at the phase shift method it is widely used in the ordinary rangefinders also in the one I used in video number 198 these devices have a laser diode which they modulate with a frequency of let's say 10 megahertz the sinus curve of this signal looks like that its wavelength is 30 meters let's assume we add a reflected signal which traveled 7.5 meters this signal will be smaller and 25% late if we were able to detect this phase difference we would be able to measure distance let's assume we can distinguish a 1 degree phase difference then we would be able to detect differences of 30 meters divided by 360 which is around 8 centimeters and the maximum distance we would be able to measure is 30 meters after the 30 meters the wave starts again and we do not know if we are between 0 and 30 or between 30 and 60 meters the measured phase would be the same this is already a reasonable system a lower frequency would increase the range and decrease the accuracy and a higher frequency the contrary however this cheap rangefinder from China can measure 40 meters with a precision of a few millimeters how come the device uses first a low frequency let's assume 10 megahertz to decide which approximate distance the target is then it increases the frequency let's say to 100 megahertz and measures again now the wavelength is only 3 meters and it can measure with an accuracy of 0.8 centimeters let's assume during the second measurement it detects a phase shift of 50 degrees then the distance could be 40 centimeters 3 meters and 40 centimeters 6 meters and 40 centimeters 9 meters and 40 centimeters and so forth because we know that it has to be around 7.5 meters we know it is exactly 6 meters and 40 centimeters because this is the closest multiple of 40 centimeters this is precisely how these rangefinders work as you see here on the oscilloscope they send a few signals with increasing or decreasing frequencies and do the measurements and calculations mentioned before this can be done with cheap components because we do not need to deal with picoseconds and still get high accuracy unfortunately this principle is rather slow my device had a minimum measuring time of 300 milliseconds but for those handheld devices this is okay by the way for my German speaking viewers I leave a link to a video where this principle is demonstrated now to the supreme discipline measuring the time of flight here we use a so called time to digital converter it has to be able to measure very short times as you see in the chart taken from the application note of the TDC - GP 22 this chip has a basic accuracy of 90 Pico seconds in which light travels around 10 millimeters incredible the block diagram shows how a device works a microcontroller pulses a laser diode which sends a laser beam towards the target a part of this light is reflected directly to a detector which stops the first counter which is used as a reference the rest of the beam travels to the target and back and is detected by a fast and very sensitive photodiode its signal stops the second counter and the chip can calculate the distance I have two of these modules here the TF mini which only uses an LED and not a laser diode and it has a specified range of 12 meters the second is from Garmin and uses a laser diode its specified range is 40 meters the TF mini costs around $40 and the armen around 150 including shipping not exactly cheap devices so let's test them I use for both a ready-made library to avoid reading the manuals like that I'm not sure if the configuration used for this test is the best available but it is how most of us would do and this is what counts on this channel after my disappointment with my last laser rangefinder I'm a little skeptical for the maximum distances again I use our big garage as a test location like that we have no sunlight but the reflection of the walls is also not ideal gray concrete let's start with the distance of about one meter a remark at the beginning I'm not interested in the absolute precision of the measurements also because I'm not equipped to do such analysis I use the rangefinder as a reference it shows one meter the Garmin does not give a precise reading it fluctuates quite a bit and stochastically it displays a completely wrong result I did not find out why this happens but because it is far off it would be easy to remove it the TF mini shows a constant value of 96 centimeters the three centimeters can be attributed to the difference in the reference plane between the rangefinder and the TF mini I did not adjust it properly the next stop is at 5 meters the Garmin still creates variable results and the TF mini very stable ones they seem to be a few centimeters short at 7 meters the TF mini has a little signal strength and stops to display results if I install a Styrofoam sheet in front of the grey wall the TF mini displays the distance the reflection of the white layer is better than the one of the gray concrete it also works at 10 meters but at around 11 meters it stops so it nearly reaches the specified range good because these lidar sensors do not emit any visible light I had to add a red laser diode to my setup like that I can aim much better what about the Garmin it still works at 22 meters at 30 meters and it displays sporadic results at 40 meters also here the specifications are met under good conditions of course with a white surface and low ambient light as we saw a great target can reduce the reach to nearly half by the way the Garmin has to be calibrated every 100 measurements this is built into the code and runs completely automatic next I'm interested in the speed the specifications of the TF mini say 100 Hertz and we measure 8.8 milliseconds for one measurement which is okay the Garmin needs 3.6 milliseconds for its measurement which is around the specified 270 Hertz so both reach the specified values a newer version of the Garmin lidar should be faster and need less power but it is also more expensive the lidar sensors are not only made for situations as I tested they are also made for rotating devices this is why I started to print some parts for my TF mini for a 360 degrees rotational lidar but because I got a ready-made 360 degrees lidar I stopped this project and will mount the TF mini or the sharp sensors onto a servos you see the results in this chart it detects obstacles in an angle of up to 180 degrees like that I can now replace the three ultrasonic sensors on my robot with one lidar and it has a longer range which can be important for this application if your device moves ask and because I can turn the robot on the spot I even do not need a 360 degree slide or if I want to look back I just turned the whole robot a viewer wrote a library to use a stepper motor instead of a servo which is probably a more reliable way you find a link to this sketch in the description summarized we know now at least four different methods to measure distance using light measuring the signal strength of the reflected light beam works fine on short distances it is cheap but not very accurate it is a good solution for slow robots triangulation is good for short distances it is fast and precise but I did not find a ready-made sensor which uses this method face detection is widely used for the handheld rangefinders it is very precise for distances below 100 meters but because it has to perform a few consecutive measurements to get accurate distances it is quite slow and not useful to build lidar systems measuring time of light is the newest technology it became cheaper and we already get small and cheap sensors like the VL 53 L o X this shape works up to about one meter and is quite slow so it cannot be used for a lidar the newer BL 53 L 1 X should double the range and increase the sampling speed to 50 Hertz this could become a good solution for short-range ly doors if the pricing is okay so far I did not see any breakout ports using this tube the TF mini is located between the cheap very short range ships and the expensive Garmin device it fits when you need exact distances in a fast pace it is also much smaller than the Garmin it could be a good replacement for ultrasonic devices where you need a waterproof solution for example to monitor all sorts of liquid tanks bigger devices like the one from Garmin covers a range of up to 40 meters they are fast enough to be used on fast moving vehicles but cost more than $100 if you want to secure your fast drone or if your autonomous buggy has a big engine it is probably the right solution if you spend even more you get devices which already have the 360 degrees movement integrated fortunately also these sensors get cheaper I just got a brand new one which costs less than $100 but this is stuff for another video a unique device is the long distance range finder which claims to be able to measure up to 1500 meters do you believe that also stuff for another episode you see this technology is not cheap this is why I want to thank all my supporters on patreon and the viewers using the affiliate links in the description for their shopping without you it would be difficult for me to do what I do now bye

Channel: Andreas Spiess

Views: 184,668

Keywords: iot, ttgo, stepper motor, triangulation, eevblog, how to, Garmin lidar, stepper, greatscott, smart home, simple, esp32, guide, beginners, LIDAR, lorawan, Rangefinder, esp8266 datasheet, electronics, do-it-yourself, time-of-flight, arduino, Sharp, Lidar Arduino, TFmini, wemos, esp32 datasheet, tof, diy, hobby, things, esp32 project, servo, TFmini LIDAR, LIDAR SENSOR

Id: ddxguAzzzJE

Length: 16min 30sec (990 seconds)

Published: Sun Jun 10 2018