object recognition using ultrasonic sensors

Hey everyone!

I’m having some trouble with an algorithm that I need to write for my studies and need some help.

So, basically I have 2 ultrasonic sensors mounted on servos, that are measuring the distance between them and an object. Now using the distance from the measurments, it needs to return an information if the object in front of them is either: a wall, a corner or a pole.

The wall and corner part are easy enough, but the pole is where the fun begins.

Has anyone an idea, how to approach this problem?

Also Thanks in advance for your help :slight_smile:

If you know the difference between a wall and a corner, then the only thing left is a pole.
So, if not wall and not corner, then pole?


Pretty much, yes.

Sorry if that was kinda confusing, but every time I tried asking someone about my problem, they wanted to know, what kind of "object" recognition I had in mind with some details.

One approach to solve the recognition problem is to make lots of distance measurements on several known objects, and study those measurements carefully.

If there is a recognizable pattern in the distances that can be uniquely associated with one of the objects, then figure out how to identify that pattern in the raw measurements.

Keep in mind that measurement noise, different environments, objects not included in the original study, and different points of view will all confuse the issue.