A noise reduced car communication system

In the communication system used in car, voice signal from the driver or passenger acquired by a microphone is contaminated by background noise, such as noise from the engine and audio signal from the car entertainment system. As a result, the person with whom the driver or passenger is communicating constantly hears the loud background noise. This may greatly impair the understanding. To cope with this problem, we propose an adaptive noise reduction solution. The background noise acquired by the microphone can be suppressed by applying the adaptive filters whose coefficients are adjusted by a Least Mean Square (LMS) algorithm. We build up an experimental system to test the effectiveness of our solution. In our experiment setup, the noise sources are two loudspeakers with each of them playing simulated engine noise and audio signal from the car entertainment system respectively and other loudspeaker is used to generate wanted voice signal. A microphone acquires the noise-contaminated voice signal. An Analog Device's SHARC EZ-KIT LITE is used to implement the adaptive noise reduction algorithm. After setting up the system and fine-tuning the parameters, we measure the effectiveness of our solution, and it shows that more than 20 dB reduction can be achieved.