Abstract
Active noise control (ANC) has been widely utilized to reduce unwanted environmental noise. The primary objective of ANC is to generate an anti-noise with the same amplitude but the opposite phase of the primary noise using the secondary source. However, the effectiveness of the ANC application is impacted by the speaker's output saturation. This paper proposes a two-gradient direction ANC algorithm with a momentum factor to solve the saturation with faster convergence. In order to make it implemented in real-time, a computation-effective variable step size approach is applied to further reduce the steady-state error brought on by the changing gradient directions. The time constant and step size bound for the momentum two-gradient direction algorithm is analyzed. Simulation results show that the proposed algorithm performs effectively in the time-unvaried and time-varied environment.
Introduction
Active Noise Control (ANC) systems function by generating an "anti-noise" signal—a wave with identical amplitude but opposite phase to the primary noise. A significant practical challenge in ANC implementation is speaker output saturation. When the required anti-noise exceeds the physical capabilities of the secondary source, performance degrades, and standard adaptive algorithms like Filtered-X Least Mean Square (FXLMS) may diverge.
Existing solutions, such as the Two-Gradient Direction (2GD) algorithm, address saturation by switching the update direction when limits are reached. However, this often introduces weight oscillations at the boundary, leading to steady-state error, or requires excessively small step sizes that slow down convergence in dynamic environments.
Proposed Methodology
The proposed 2GD-Momentum algorithm with variable step size (VSS) enhances the standard FXLMS framework with three integrated components:
- Two-Gradient Switching: The algorithm monitors output power. Within limits, it minimizes error using standard gradients. Upon exceeding limits, it switches to a secondary gradient that steers weights back toward the constraint boundary.
- Variable Step Size (VSS): To mitigate weight "vibration" at the saturation threshold, the step size is automatically reduced upon hitting the boundary, ensuring a smoother transition and lower steady-state error.
- Momentum Factor ($\kappa$): A momentum term is introduced to counteract the reduced convergence speed caused by VSS. By utilizing the "inertia" of previous updates, the system maintains rapid adaptation, which is vital for time-varying acoustic paths.
Experimental Evaluation
The algorithm was evaluated using simulations based on real-world acoustic paths measured from an air duct (16 kHz sampling). Noise sources included broadband (200–800 Hz) and White Gaussian Noise.
Key Results
- Stability and Convergence: As shown in the weight magnitude plots, the 2GD-Momentum method exhibits significantly lower fluctuation compared to standard 2GD and rescaling techniques. In time-varying scenarios where the environment changes suddenly during operation, the proposed method achieves faster recovery to the new optimal operating point.
- Spectral Performance: The algorithm successfully attenuates noise in the target band without the frequency spikes often seen in unconstrained FXLMS under saturation.
Conclusion
The 2GD-Momentum algorithm effectively balances the trade-off between output constraint enforcement and convergence speed. By incorporating momentum and variable step sizes, it achieves a stable, sub-optimal solution at the hardware limit. Future work will focus on real-time hardware implementation on DSPs and integrating online secondary path modeling to handle more complex real-world dynamics.
Summary updated