Experimental results for multichannel feedforward ANC with noninvasive system identification

Area of Science:

• Acoustics
• Signal Processing
• Control Systems Engineering

Background:

• Active noise control (ANC) systems traditionally struggle in environments with time-varying transfer functions.
• Adapting ANC algorithms to dynamic conditions is crucial for effective noise reduction.

Purpose of the Study:

• To experimentally validate ANC algorithms capable of operating effectively despite significant changes in environmental transfer functions.
• To assess the performance of adaptive ANC with and without real-time system identification in a realistic room setting.

Main Methods:

• A multichannel ANC system was implemented in a reverberant room (1.4 x 2.4 x 2.4 m3) with a seated occupant.
• Six microphones defined a quiet zone near the occupant's ears, with four secondary sources driven by a single reference signal.
• Experiments compared an adaptive cancelling algorithm with static models against one integrated with a noninvasive real-time system identification algorithm.

Main Results:

• Static model ANC achieved over 23 dB sound pressure level reduction when the occupant remained still.
• Adaptive ANC with real-time system identification achieved approximately 18 dB reduction, robust to occupant motion.
• Performance was evaluated over the 50 to 1000 Hz frequency range using an ideal reference sensor.

Conclusions:

• Real-time system identification significantly enhances the robustness of adaptive ANC in dynamic environments.
• The validated algorithms demonstrate practical applicability for local broadband noise control in occupied spaces.
• ANC systems can be adapted to accommodate occupant movement, improving real-world performance.