Coin recognition system based on a neural network

The purpose of this paper is to design an electronic system which can identify coins by detecting the sound generated when they hit a hard surface, using a neural network. Generally, coin identification in vending machines is done using magnetic or optical methods. This paper focuses on the acoustic method, in which coin recognition is based on the detection of the coin's natural frequencies. The frequencies of these vibrations depend on the object's properties (mass, shape, material type), and remain the same as long as these properties do not change, thus being used as acoustic fingerprints. Also, this method permits recognition of fake or deteriorated coins, because they have different properties. The principle applied in this paper can be used for the recognition of numerical sequences produced by other objects.