Robotshop listed Lidar Lite V4 says it is LED, but a review I read said it was laser. Garmin site lists the unit for $129 USD, Robotshop is selling for $79 CAD. So something is …wrong.
How good is the Lidar Lite V4 (LED or laser) compared to TF-mini, for long range detection of moving objects?
1 Like
Hey @queenidog,
Welcome to the RobotShop community!
Well, I’ll assume Garmin know their own their device and are not lying. The following is taken directly from the Garmin website (page for the Lidar Lite V4 LED):
Are you sure you are not confusing the V3 (which is about that price) with the V4 (which seems about 60 USD everywhere (resellers), including on RobotShop’s US website)?
Well, they are both good devices but they do have some differences:
- The range is similar enough (10 m vs 12 m), slight (potential) advantage on the TF-Mini there, but it really depends on how fast that maximum range drops for each at lower reflectivity / angled surfaces.
- Lidar-Lite V4 has lower power consumption (active=85 mA peak) vs TF-Mini which has 120 mA average.
- The Lidar-Lite V4 had double the measurement rate (200 Hz vs 100 Hz), which would be advantageous at long distances. That being said, both seem to be presenting their “best case scenario”. Expect lowest rates for both.
- The Lidar-Lite v4 seems to be mostly directed for indoor use whereas the TF-Mini has a 70klux immunity. Direct sun light is ~100kLux so that is still not enough for mid-day outdoors (without shading), but it may perform better outdoors.
- The accuracy on the Lidar-Lite v4 is much better. It has about 0.5% accuracy (so, about 1 cm @ 2 m of range) whereas the TF-Mini has 1% up to 6 m and then 2% at 6-12 m, which if you are looking at longer ranges specifically, that is 4 times worse. At 10 m, 2% = ±20 cm in range measurement which is quite significant!
- Oh, and finally, the Lidar-Lite v4 can communicate and be updated by wireless (ANT+), so that’s a big bonus. The TF-Mini has no wireless interface and thus must be hard wired to your microcontroller. This could be very helpful when placing the sensor on a moving part (such as a “turret” for sensing from many angles, no slip ring needed!) or when wires would be inconvenient!
All things said, neither are particularly great for detecting moving objects. They both are single point range detection which is more suitable for detecting large obstacles or walls/floors/ceilings/etc.
I hope this helps!
@queenidog
Why not use ultrasonic rangefinders? The car industry uses them for park-assistance. Not sure about the max. range. but 20m would be possible I guess. Ultrasonic works in any environment (regarding ambient light conditions) which is a big advantage.
2 Likes
They can work really well indeed. They do have a few caveats though.
Since they use sound, they have a limited rate based on what they detect (lidars do too, but light is MUCH faster than sound). If trying to detect objects near the maximum range of the sensor, you’ll get a slow rate when there is nothing to bounce the signal back (timeout).
Also, the active detection signals are typically much more divergent (large cone-ish shapes) and therefore need to be far from the ground for a 10 m (or more) range, otherwise you can get echos from the floor being detected as an object (heavily depends on sensor here, some may handle echos but most don’t).
And while they will not be affected by light, as mentioned above they have a wide beam so it is difficult to detect individual objects precisely - especially a long ranges. They are far better suited at range finding for obstacles as a whole, hence why they are used in car park assist (and similar task) as they typically won’t miss a smaller obstacle due to the size of their detection “cone”.
It really depends on what @queenidog is trying to do specifically. If detecting objects at long range is the plan, I’d say mm-wave radar might be better actually! Maybe something from RB-Omn.
Sorry for the long, long reply…was working on other projects but this one is coming around again. I want to make a vehicle counter that counts vehicles as they pass by on the highway. Speed liimit is about 60 kph, but of course, people will drive slower or faster. I’ve made prototypes using US, LIDAR, laser in my shop where I can detect my hand flashing by, about 10 feet away. For the highway I need about 30 feet (10 meters) to get both lanes one going one way, the other opposite way. The devices in my Ford Escape seem to work very well in this applicaiton, detecting anything that I could be backing into and detects moving vehicles perpendicular to my car movements. THAT’S what I want!! I purchased a few US sensors from Robotshop. I tried a prototype US (quality one from robotshop) but there’s so much “noise” with US that needs filtering. I found LIDAR and laser (in my shop to be more “cut and dry”. The signal is there or it isn’t. Any idea what the automotive industry uses?