在高噪音的工作廠區中，噪音來源往往來自不同類型的設備與作業活動，例如大型機械運轉產生的低頻震動噪音、高速切割機具的高頻尖銳噪音、氣動設備與壓縮機的間歇性脈衝噪音，以及人員交流或警示信號的環境音。由於這些噪音類型不同、頻率範圍廣且可能互相干擾，傳統的降噪技術難以全面有效應對。透過 AI 模型學習各類噪音型態，針對多種異質噪音源進行即時降噪，無需配戴耳機即可顯著降低環境噪音對作業人員的影響。在某些場域中，可能需要根據不同噪音類型與來源動態建立多個靜音區，以實現精準降噪與區域化調整，多噪音源主動噪音控制技術因此成為解決此類問題的關鍵方案。本研究主旨在針對多種異質噪音源進行即時降噪，使作業人員無需配戴耳機即可顯著降低環境噪音對其的影響。因此，本研究將設計頭枕型ANC系統，當使用者站在頭枕前，系統即可依所設置的參考麥克風接收環境噪音信號，系統的次級喇叭可產生反噪音信號進行破壞性干涉來抵銷噪音。 In high-noise industrial environments, noise sources often originate from various types of equipment and operational activities. Examples include low-frequency vibration noise generated by large machinery, high-frequency sharp noise from high-speed cutting tools, intermittent impulsive noise from pneumatic devices and compressors, as well as ambient sounds such as human communication or warning signals. Because these noise types differ in nature, cover a wide frequency range, and may interfere with one another, traditional noise-reduction techniques are often unable to effectively address them comprehensively. By enabling AI models to learn different noise patterns, real-time noise reduction can be achieved for multiple heterogeneous noise sources, significantly reducing environmental noise exposure for workers without requiring the use of headphones. In certain environments, it may also be necessary to dynamically create multiple quiet zones based on different noise types and sources to achieve precise, localized noise reduction. Therefore, multi-source active noise control (ANC) technologies have become a key solution to such problems. The objective of this study is to achieve real-time noise reduction for various heterogeneous noise sources so that workers can experience significant noise relief without wearing headphones. Accordingly, this research proposes the design of a headrest-type ANC system. When a user stands in front of the headrest, the system’s reference microphones capture the environmental noise signals, and the secondary loudspeakers generate anti-noise signals that destructively interfere with the noise, thereby achieving effective noise cancellation.