How does a lidar work?

[Electronics4you]

Hi all,

how does a lidar work. I have heard that it pulses laser by 1000Hz, and then measuring the time it takes to come back, but that will need incredible high speed computering. There might be a better explanation.

Electronics4you

 

[jbeng]

Actually, time-of-flight (TOF) is how the police lidar works. The unit sends out a pulse of laser light which is reflected back from the target. The unit knows when it sent the pulse and senses the reflection and calculates the distance to the target based on the TOF. It repeats this after a known period of time and calculates the difference in distances. The distance difference divided by the time between measurements equals speed. Doing a bit of math tells us that light travels approx 9.84 x 10e8 feet per second (186,282mi/sec x 5280ft/mi) . This means that it will make a half-mile round-trip (1320ft x 2, a quarter mile to the target and back) in about 2.7 microseconds, which is well within the processing capabilities of high-speed computers these days.

Here's a link: https://auto.howstuffworks.com/question396.htm
JB

 

[Oznog]

This means that it will make a half-mile round-trip (1320ft x 2, a quarter mile to the target and back) in about 2.7 microseconds, which is well within the processing capabilities of high-speed computers these days.

That IS pretty high speed. That's only 3.7 cycles of a 10MHz clock for a pretty long trip. I'd wondered about making something like this and quickly saw it was impractical to get accuracy of a few feet due to the quantization errors alone. Seems like a very special input stage would be necessary rather than counting clocks. Even more of a question is what kind of sensor has a response time that small?

 

[akg]

That IS pretty high speed. That's only 3.7 cycles of a 10MHz clock for a pretty long trip. I'd wondered about making something like this and quickly saw it was impractical to get accuracy of a few feet due to the quantization errors alone. Seems like a very special input stage would be necessary rather than counting clocks. Even more of a question is what kind of sensor has a response time that small?

may be they are using interferometric method , or variable freq laser .?

 

[crust]

I haven't posted in a while .... Anyway, it is true that the actual number of clocks is pretty small. In fact if you consider the pulse to be an impulse, and send it at time t, then every one will come back at some time t' but you will measure it (in clock ticks). The main problem is that you don't know whether it was close to the beginning of the clock period or the end. There are plenty of tricks to get around this problem such as histogramming, but consider this rather simple approach:

The receiver will assert some signal when the laser pulse gets back. Now even if the receiver only runs with a 100ns period (10MHz). We could get
a very accurate measurement if we knew where within the period the pulse arrived. The idea is that you take the signal and route it to many
flip-flops each one gated by a clock that is slightly delayed with respect to its peer. For instance, Bit 0 is at t0, Bit 1 is at t0 + 5ns etc. You do this up to half the period, and then flip the clock and now you have the full
period covered and now when you latch the bits, some will be 0s and at
some point you will see 1s. Knowing what the delay is, you can then very
accurately determine the pulse measurement.

Another approach is to make the pulse operate at a faster frequency than
the receiver. The pulses are received and accumulated by a histogram.
The mass-center then tells you where the distance was (not exact, but for
speed you only need x'-x) and you do it again and again and the hist. will
"translate" giving you the velocity.