The object discrimination system using a neural network with inputs for distance and sensitivity information of an ultrasonic sensor

Proposed a discrimination system using a neural network with inputs for distance and sensitivity information of an ultrasonic sensor. This system consists of the following three elements: (1) ultrasonic sensor receiving and transmitting ultrasonic wave; (2) transducer that changes the output signal of sensor to distance and sensitivity information; (3) neural network for object discrimination with inputs for distance and sensitivity, with outputs corresponding to each object. We confirmed the validity of the system by experiments using different kinds of objects, objects with different grain size, and objects with different surface roughness. The proposed discrimination system can be applied to industrial processes.