透過您的圖書館登入
IP:18.221.58.143
  • 學位論文

利用智慧型手機並結合語音辨識進行汽車防盜系統遙控

Voice-recognition-based Remote Control Using Smart Phones for Car Security Systems

指導教授 : 蔡偉和
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


本論文嘗試利用智慧型手機取代傳統遙控器以進行汽車防盜系統的控制。基本操作是藉由藍芽無線通訊來傳遞控制指令,例如上鎖與開鎖等。但為了避免因手機遺失而造成進一步車輛遭竊的重大損失,我們透過語音身分確認來達成更安全的防護。同時,這種使用智慧型手機來控制汽車防盜系統的另一項好處是可讓多位使用者操作,例如全家人都可持手機遙控,省去複製傳統遙控器的成本與麻煩。另一方面,考慮使用者以語音進行遙控時,可能為了避免他人聽見或側錄而以輕聲細語方式說話,這種輕聲細語常因氣音過多而造成錄音的量化飽和,即所俗稱的「爆音」。我們發現當爆音發生時,語者確認系統的效能將明顯下降。本論文因此進一步提出透過爆音音框刪去方法來改善爆音所造成語者確認效能下降的問題。經由實驗結果證明,本論文所提出的方法可有效降低輕聲細語下之語者確認系統的「等錯誤率」。

並列摘要


This thesis proposes a novel user interface for car security systems via smart phones in an attempt to replace remote controllers. The basic strategy is the use of Bluetooth to transmit the control commands, such as lock and unlock from a smart phone to the in-car security system. To prevent from the imposters arising when the smart phone of a vehicle owner is stolen or lost, we propose using speaker verification to provide double check for a user’s identity. Another advantage of such a smart-phone-based control for car security systems is to support multi-user without extra cost, which eliminates the need from copying a remote controller. On the other hand, when a user speaks to the system, he/she may use whispering to avoid his/her voice from being heard or recorded. However, it is found that whispering may contain aspirated speech, which is prone to arising quantization saturation or termed “clipped speech” during the whispering is recorded. The clipped speech could degrade the performance of speaker verification system. The thesis proposes a method of “frame pruning” to alleviate the problem of clipped whispering speech for speaker verification. Our experiments showed that the proposed method can reduce the equal error rate (EER) noticeably when dealing with clipped whispering speech.

參考文獻


[5] Huanbing Gao, Liyan Yuan and Tao Wang, “The Family Safety Monitoring System Based on Ethernet and Bluetooth,” World Congress on Intelligent Control and Automation, China, July 2010, pp. 4281-4285.
[6] Xing Fan and John H. L. Hansen, “Speaker Identification Within Whispered Speech Audio Streams,” IEEE TRANSACTIONS ON AUDIO, SPEECH, AND LANGUAGE PROCESSING, vol. 19, no. 5, July 2011, pp. 1408-1421.
[7] Wang Yanlei, Zhao Heming, Gu Xiaojiang and Gong Chenghui, “A Study on Speaker and Session Variability in Speaker Recognition of Chinese Whispered Speech,” International Conference on Industrial Mechatronics and Automation, 2010, pp. 292-295.
[8] Qin Jin, Szu-Chen Stan Jou, and Tanja Schultz “WHISPERING SPEAKER IDENTIFICATION,” International Conference on Multimedia and Expo, 2007, pp. 1027-1030.
[9] Taisuke Itoh, Kazuya Takeda and Fumitada Itakura, “ACOUSTIC ANALYSIS AND RECOGNmON OF wmSPERED SPEECH,” IEEE, 2002, pp. 389-392.

延伸閱讀