0
\$\begingroup\$

I am currently challenged with the following problem:

I am given an autonomous robot which uses the HC-SR04 ultrasonic distance sensor to determine its distance to the robot driving in front. Depending on the angle the robot driving in front has to the robot driving behind, some sound waves seem to be reflected in such a way that they are not detected by the receiver of the robot driving behind. After a quick Google search this problem seems to be known as specular reflections. Does anyone know a good way of improving the situation? I.e. I want more reflections to be detected by the receiver of the robot driving behind.

I was thinking: Maybe it is a good idea to use LIDAR instead of ultrasonic-distance sensors. But won't laser beams suffer from the same problem? There seems to be an algorithmic approach of improving the ultrasonic-distance measurements using a so-called Dempster-Shafer. This appears to be highly complicated, however. Plus, I am using a Kalman-Filter to improve my sensor-readings already. So I am not sure if the Dempster-Shafer would collide with my Kalman Filter.

Anyways, I'd be so grateful for any help! I am willing to read about this topic as well - if you know any good sources, please share!

\$\endgroup\$
  • \$\begingroup\$ Can you modify the robot in front, or only yours? \$\endgroup\$ – Jack B Feb 8 at 11:41
  • \$\begingroup\$ I can modify both robots! Are you thinking about modifying the surface of the robot in front somehow? \$\endgroup\$ – user503842 Feb 8 at 11:42
  • 1
    \$\begingroup\$ Yeah. I don't know much about ultrasound sensors specifically, but generally speaking if you have a specular reflection and you want a diffuse one you can do that by roughening the surface. You want it to be "rough" on a wavelength scale, so a bit less than a milimeter for ultrasound. \$\endgroup\$ – Jack B Feb 8 at 11:54
  • 1
    \$\begingroup\$ This isn't an EE question - try asking on a different forum (not sure what). \$\endgroup\$ – Andy aka Feb 8 at 11:57
  • 1
    \$\begingroup\$ In past decades I assisted on robot projects using both 128*128 images, and a single scanning ultrasonic (Polaroid Camera setup, from TI). We were troubled by CORNER REFLECTIONS, where the side-lobes produced unwanted false-returns. In largely empty rooms, the robot was halted because of objects nearby yet 30 --50 degrees off-axis. So beware. \$\endgroup\$ – analogsystemsrf Feb 8 at 12:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.