Enhanced multi-channel active noise control system and real-time implementation for infant incubator
The active noise control (ANC) system is implemented in the infant incubator and confirmed to improve environment effectively for infants. However, the traditional ANC systems cannot be applied directly to incubators without modification due to complicated noise models in the neonatal intensive care unit (NICU). It has multiple unexpected noise sources both inside and outside the incubators. The thesis proposed a system which combines the feed-forward and feedback ANC systems with multiple reference microphones, generate a larger zone of silence, offer deeper noise attenuation, and operate in complex noise environments, i.e. multiple unknown various noise sources. Moreover, the on-line secondary path modeling algorithm is also designed and implemented to solve the time-varying secondary path problem.This thesis implemented the enhanced active noise system for infant incubators with noise source detection for improving the ANC system's performance. After detect the noise source direction select the relative closer reference microphones and preprocess the received signal can solve the unknown noise source problem. The hybrid ANC algorithm can cancel both inside and outside noises for incubator. In order to track the time-varying secondary path, on-line secondary path training is also designed and implemented. In infant incubator application, a single-channel hybrid active noise control (MHANC) system is implemented for such hospital noise environment. Studied a real time MHANC system implementation is being executed a TMS6713 Quad Processor Board from the HIRATSUKA Engineering Co., Ltd.