Were you ever able to connect via I2C? Were you able to use the acconeer exploration tool to connect or able to use pyserial?

To measure speed using a pulsed mode radar (if you don't access to Doppler information...) you would start a timer at the instance of your transmit pulse, and then stop the timer when you get a return pulse - how long it takes for the pulse's round trip is directly related to the distance. They are already doing that for you in the distance measurement routine. You would save that distance in a variable, then take another distance measurement and if it has changed, either coming closer, or getting further away, you would want to know the time it took for it to move that distance - that would be the speed - you are basically measuring the distance of travel over time.

You would want to take a series of those measurements and average them a little to weed out the fact the target will likely not be a perfect reflector - One reading might be of the front bumper, the next you might get a stronger return off the rear wheel - those kinds of things will add "noise" to your data you would need to account for.

That's how police LIDAR works, but its easier because a laser beam is still relatively narrow a distance and you don't have near the problems with the beam reflecting off of random surfaces - unless the officer has a really shaky hand. But they have software that is averaging and throwing away any wild returns, it needs to send a series of pulses to do this, but the clock rates of course fast enough the operator will just get a solid speed reading and distance, and the speeder will get a ticket before they can blink. The problem being with that is they can't be moving. Its stationary operation only. But they can use the distance measurement the LIDAR provides to check for tail gating, accident scene measurements, etc. Its also very hard to defeat outside of methods we wont talk about here... Radar detectors are absolutely useless against LIDAR because by the time you hear an alert, the nice LEO has your speed, and if you are breaking the law, good luck. All because it is an optical, line-of-sight dependent device that isn't spraying the air waves with RF - With radar, the ball is always in the violator with a radar detectors court due to 1/R^4 > 1/R^2

Radio waves spread a lot at distance, and by looking at the measurement distance of this thing, of only 2 meters, its kind of apparent it wouldn't be very useful for speed radar. You might be able to make some sort of dielectric lens, like a on a real police radar. If you've ever noticed the antenna of one of those it will have lens looking plastic dome on it, a bit like a can of soup with a spot light lens stuck in the end of it. That is a dielectric lens, it actually focuses the beam of RF from horn antenna inside the can just like an optical lens would focus a light beam from a light bulb for a narrower beam width at distance. I'm sure there is some sort of modeling software out there that could help design the lens, and you would probably need it - but at 60GHz, that is roughly twice as high in frequency as Ka band police radar, so the lens for this would need to probably be roughly half the dimensions?

I haven't read the datasheet for this thing, but I imagine it was intended for blind spot detection or collision avoidance more than long distance speed measurement.

Big radars, like weather or military types would be using the Doppler information of the return pulses for velocity information, as it would be a direct "measurement" rather than a distance over time "calculation" and much more reliable data. In fact, that is what police speed radar uses, the Doppler shift caused by the motion of your car in relation to the transmit frequency, which is constant, or constant wave (CW) mode because they don't really car about distance information and it's a lot cheaper than pulsed mode radar. The other advantage to CW Doppler police radar is they can be in motion and still clock violators, by taking into account the police cruisers speed via Doppler and throwing it out, leaving only target speeds. Again, not having read the datasheet or "SDK" - Maybe they give you a Doppler output? Then it would be a simple matter of doing the math to correlate Doppler shift to velocity

The only real advantage for a pulsed mode police radar would be a lower probability of intercept for those using ECM - i.e. a radar detector... In other words, it would be all the more difficult to detect a police cruiser running pulsed radar vs. CW radar. Again - 1/R^4 > 1/R^2 is a win pretty much every time for the radar detector and anything that can help offset that is an advantage to the radar and its operator....

Unfortunately not, the only examples in the SDK are a distance detector, and how-to's on using the IQ, power bin, and envelope services. Hopefully Acconeer continues building on the SDK to produce more application-esque examples.