| In neonatal intensive care units (NICU), infant incubators are the places where premature babies are nurtured for their complete development. The noise in NICU environment has lifelong adverse consequences to the NICU graduates. This thesis develops active noise control (ANC) systems for infant incubators in NICUs. Several ANC algorithms are evaluated for the complicated problem of noises generated from different sources both inside and outside the incubator. A multiplechannel hybrid ANC system is developed effectively to cancel both inside and outside incubator noises. Computer simulations and real-time experiments are conducted to evaluate the performance of the feedforward, adaptive feedback, pseudo, and hybrid ANC systems using the recorded inside and outside incubator noises.;The ANC system for the infant incubators is further enhanced by integrating a cry detection algorithm. Sound recognition techniques are studied and most suitable ones are used to detect the cry signal embedded in noises such as the NICU noise, and speech and footsteps of the people moving inside the NICU. |
