Problem with the guardrail radar detections using Automated Driving Toolbox

Hello all,

I am trying to plot two graphs such as radial doppler velocity vs azimuth/elevation using a driving scenario. The problem is when I add a guardrail to the road, the elevation vs radial doppler velocity graph is not meaningful like the radial velocity vs azimuth. I have tried with two different roads and the graph is problematic with a guardrail. Please see the images below. Whereas if I remove the guardrail, I get sensible graphs. Could anybody else please try and let me know if the guardrail has any bugs?

In the first figure(driveScenario.fig), the general setting of the scenario is shown.

In the second figure(ele_fig.fig), the elevation vs doppler velocity just before the guardrail is shown(the weird case as the radar data scatter should be more concentrated)

In the final figure, the azimuth vs doppler velocity after crossing the guardrail and when just before stopping is shown(this is much better)

I have also simulated with different road track. Again the one without a guardrail seemed to be better. Hence my question.

It is adapted from the original

https://www.mathworks.com/help/driving/ug/model-radar-sensor-detections.html

Thanks for your time.

Answer 1

[Serge]

Hi Balaji, the plots you are getting seem as expected. The spread in the Doppler elevation figure is on the Doppler axis and it is expected as your radial velocity is different at different angles between the radar and the guardrails. When you run the simultion, you can see that the spread increases the closer you are to the guardrail as you will have different points of reflections, each with a different angle. I don't think I noticed any bug.

Also, you should be aware that when guardrails are removed the only reflections that you will be getting is coming from the target in front of your radar, whereas with the gaurdrail present, you will be getting reflections from your target and every reflection point on your guardrail.