Design and application of an automated baby cradle with cry detection

This paper is about the creation of an artificial neural network (ANN) in MATLAB to analyze the features extracted from calculating the mel-frequency cepstral coefficients (MFCC) of a raw audio data to determine if it contains an infant crying or not and interface it to a motor using a PIC microcontroller. A DC motor is used to enable the cradle to rock when a cry is detected, and the UART function of the PIC microcontroller is utilized to be able to interface the motor to MATLAB. The ANN uses multiple infant crying sounds, as well as non-crying sounds, to create a sample training set with corresponding targets that determine whether the sound is a cry or not. It uses relevant concepts heavily utilized in speech recognition for the design of the infant cry recognition with some modifications to fit the application better