LIDARs have the advantage that they allow detecting solid objects that have not been detected by a vision-only system. For example, some time ago, a Tesla crashed into an overturned truck, likely because it didn't detect it as an obstacle.

A system that's only based on cameras is only as good as its ability to recognize all road hazards, with no fall back if that fails. With LIDAR, the vehicle might not know what's the solid object in front of the vehicle using cameras, but it knows that it's there and should avoid running into it.

▲

sandworm101 5 days ago | parent [-]

Solid objects that arent too dark or too shiny. Lidar is very bad at detecing mirrored surfaces or non-reflecting structures that absorb the paticular frequency in use. The back ends of trucks hauling liquid are paticularly bad. Block out the bumper/wheels, say by a slight hill, and that polished cone is invisible to lidar.

▲

bayindirh 5 days ago | parent | next [-]

Add one or a couple of RADAR(s), too. European cars use this one weird trick to enable tons of features without harming people or cars.

▲

UltraSane 5 days ago | parent | prev [-]

LIDAR works be measuring the time it takes for light to return so I don't understand how a object can be too reflective. Objects that absorb the specific wavelength the LIDAR uses is an obvious problem.

▲

sandworm101 5 days ago | parent [-]

Too reflective, like a flat mirror, will send the light off in a random direction rather than back as the detector. Worse yet, things like double reflections can result in timing errors as some of the signal follows a longer path. You want a target that is nicely reflective but not so shiny that you get any double reflections. The ideal is a matte surface painted the same color as the laser.

	▲	UltraSane 5 days ago \| parent [-]
		Ah it relies on diffuse reflections to guarantee some light returns to the sensor but specular reflections mean none is returned. This is a good example of why sensor fusion is good.