A new method for shape recognition of a pillar-like object using a 1-D ultrasonic sensor array and GA

We propose a new method for shape recognition of a pillar-like object using a 1-D ultrasonic sensor array and a genetic algorithm (GA). The proposed system consists of an ultrasonic sensor array, a signal processing unit, and a recognition unit. In the signal processing unit, the peak positions and their values, and the positions of the inflection points and their values, of ultrasonic waves output from the sensor array unit are acquired. These values and positions are inputted into the recognition unit, which consists of the GA, for object identification. We designed and built a prototype system, and we tested it by attempting to identify six objects, each of which was a cylinder or a rectangular prism. The experimental results showed near-perfect object identification.