Adaptive ANC algorithm for internal and external dual-channel noise

Noise control technology is a burgeoning research trend due to its significant impact on speech signal quality and daily life. ANC technology is prevalent for its effectiveness against low-frequency and non-stationary noise. This paper revisits the evolution of ANC, from early concepts to modern algorithms like LMS, XLMS, and FXLMS, which have shown promise yet face limitations, particularly with internal and external dual-channel noise. In practical scenarios, receiving end noise comprises internal noise in speech and external environmental noise. Current ANC algorithms focus on external noise, neglecting internal noise and thus limiting performance enhancement. We introduce an Adaptive Dual-Channel FXLMS algorithm (D-FXLMS) to address this gap. This novel approach tackles both noise types simultaneously, improving overall speech quality by managing dual-channel noise at the receiver. D-FXLMS algorithm fuses dual-channel noise into two LMS cycles and then processes dual-channel noise at the same time. The experimental results show that the D-FXLMS algorithm is optimal regarding SNR, STOI and PESQ in static and dynamic environments. Finally, the research concludes with a summary of achievements and future directions, including further research on noise estimation algorithms and targeted processing for different noise environments to improve the system’s performance. The D-FXLMS algorithm presents a significant step forward in ANC, offering a comprehensive solution for dual-channel noise environments.