100年08月02日核正.替換頁 M413.944 五、新型說明: 【新型所屬之技術領域】 [0001] 本創作係有關於一種利用聲音控制漏電流遮斷器之 聲控裝置，尤其是指一種能提升特定與非特定語者之語 音辨識率，並以聲音控制切換漏電流遮斷器狀態之聲控 裝置。 【先前技術】 [0002] 目前所用漏電流遮斷器之控制方式.，均以手切換漏100 years, August 2, 2006. Replacement page M413.944 V. New description: [New technical field] [0001] This creation is about a voice control device that uses sound to control leakage current interrupters, especially A voice control device that can improve the speech recognition rate of a specific and non-specific speaker and control the state of the leakage current interrupter by sound control. [Prior Art] [0002] The current leakage current interrupter control method
電流遮斷器開關來達成控制目的。當有緊急危難的情形 發生時，藉由以手切換漏電流遮斷器開關，確保電源斷 電，避免引發電線走火或造成人員觸電等災害。然而， 在緊急危難事件發生之時，通常是相當緊急的，必須在 最短的時間内手動切換漏電流遮斷器開關，但因所處的 環境並非人人均能知曉漏電流遮斷器開關的裝設位置， 以致造成無法儘速切換漏電流遮斷器開關進而減輕災害 [0003] 又，語言是人類生活中最直接的溝通方式，因此使 用語言下指令的電子化產品，將更容易被使用者接受。 [0004] 為了使微控制器能以語音作為人類溝通的媒介，需 要語音辨識和語音合成技術來達成，其中文字轉語音系 統，係為一種用以將輸入的文字轉換為人類語音輸出的 技術。然而這些技術仍然存在一些缺點：合成出的語音 訊號不夠清楚，訊號中含有雜音，合成語音的說話速度 與音高頻率的變化並無支援，流暢性不佳，與實際真人 發音有明顯的差異。 表單編號A0101 第3頁/共12頁 [0005] 10C年08月02日修正替換頁 [0006] [0007] 此外，隱藏式馬可夫模型為語音辨識領域中常用之 語音模型。隱藏式馬可夫模型的特徵，係以統計為基礎 的方法，需要更多的計算，可實現非特定語者語音辨識 ，適合大詞彙，卻難以微處理器或晶片系統加以設計。 如上所述，如何在保留漏電流遮斷器原有架構下增 設一非接觸方式來控制漏電流遮斷器，令使用者可選擇 非接觸式來控制漏電流遮斷器，以及如何提升語音辨識 模型之效率，讓使用者依循原設定指令，正確地辨識語 者之語音，控制配電箱中之漏電流遮斷器，在危難與緊 急的環境下，關閉切斷電源的供應，以減低災害的持續 擴大為本創作所欲解決之課題。 【新型内容】 為此，本創作乃提供一種利用聲音控制漏電流遮斷 器之聲控裝置，係關於一種能提升特定與非特定語者之 語音辨識率，且能以聲控方式控制切換漏電流遮斷器之 狀態者，而此即為本創作之主要目的。Current interrupter switches for control purposes. When an emergency situation occurs, switch the leakage current interrupter switch by hand to ensure that the power supply is cut off, and avoid causing fires such as wire fire or electric shock. However, in the event of an emergency crisis, it is usually quite urgent. The leakage current interrupter switch must be manually switched in the shortest time, but not everyone can know the leakage current interrupter switch because of the environment. The position is set so that the leakage current interrupter switch cannot be switched as quickly as possible to mitigate the disaster. [0003] Language is the most direct communication method in human life, so electronic products using language instructions will be easier to use. Accepted. [0004] In order for a microcontroller to use speech as a medium for human communication, speech recognition and speech synthesis techniques are required, wherein the text-to-speech system is a technique for converting input text into human speech output. However, these technologies still have some shortcomings: the synthesized speech signal is not clear enough, the signal contains noise, the speech speed of the synthesized speech and the pitch frequency change are not supported, the fluency is not good, and there is a significant difference from the actual human voice. Form No. A0101 Page 3 of 12 [0005] Correction Replacement Page of August 2, 10C [0006] [0007] In addition, the hidden Markov model is a speech model commonly used in the field of speech recognition. The characteristics of the hidden Markov model are statistically based methods that require more calculations to enable non-specific speaker recognition, which is suitable for large vocabulary but difficult to design for microprocessor or wafer systems. As mentioned above, how to add a non-contact method to control the leakage current interrupter under the original structure of the residual leakage current interrupter, so that the user can select the non-contact type to control the leakage current interrupter and how to improve the speech recognition. The efficiency of the model allows the user to follow the original setting instructions, correctly identify the voice of the speaker, control the leakage current interrupter in the distribution box, and shut down the supply of the power supply in a distressed and emergency environment to reduce the disaster. Continue to expand the subject of the creative work. [New content] To this end, this creation provides a voice control device that uses sound to control the leakage current interrupter. It relates to a voice recognition rate that can improve the specific and non-specific speakers, and can control the switching leakage current cover in a voice-activated manner. The state of the breaker, and this is the main purpose of the creation.
[0008] [0009] 上述本創作之主要目的，是由以下之具體手段所實 現： 一種利用聲音控制漏電流遮斷器之聲控裝置，係包 括：一語音取音元件，裝設在電源開關箱之外部空間， 用來擷取特定與非特定語者之語音；一語音辨識元件， 用來辨識語音取音元件所接收到的語音訊號，其中該語 音辨識元件係透過一語音音效卡將語音信號由類比轉換 至數位信號，以進行系統之數位化處理，擷取該使用者[0009] [0009] The main purpose of the above creation is achieved by the following specific means: A voice control device for controlling a leakage current interrupter by using sound, comprising: a voice pickup component installed in a power switch box The external space is used to capture the voice of the specific and non-specific speakers; a speech recognition component is used to identify the voice signal received by the voice pickup component, wherein the voice recognition component transmits the voice signal through a voice sound card Converting from an analog to a digital signal for digital processing of the system, capturing the user
所發出的語音資料，以建構出語音特徵參數，並經一特 表單编號A0101 第4頁/共12頁 定語者語相以件與—料U者料 韻出語音辨識結果；以及一微處理器控制元件，於接 又到該°° s辨識元件所接收到之語音訊號之辨識結果， 將一驅動訊號輪出至該漏電流遮斷器，以啟斷電源開關 箱控制電路。 [_] 其巾料S者料韻S件，餘據料辨識模式 採用動態時間比對（Dynamic Time Warping)技術架 構，用於特定語者（Speaker Dependent)語音辨識， 藉由語音辨識微控制器（Speech Recognition Microcontroller) ， 提升語音辨識能力之效能。 [0011] 其中彝特定語者語音辨識元件，係根據採用隱藏式 馬可夫模塑/神經網絡混合（HMM/neural network hy brid)，用於非特定語者（Speaker Independent) 音辨識，藉由語音辨識微控制器（Speech Rec〇gni_ tion Microcontroller) ’提升語音辨識能力之效能 [0012] 依據本創作之利用聲音控制漏電流遮斷器之聲控裝 置，由於其係在不改變現有漏電流遮斷器操作模式下所 增設的選項控制漏電流遮斷器之方法並存，於該聲控誓 置未被觸發時’使用者仍可利用漏電流遮斷器原有的切 換方式進行漏電流遮斷器的控制。 【實施方式】 [0013] 為令本創作之技術手段、創作目的及達成功效有更 完整及清楚的揭露，茲於下詳細說明之，並請一併參閱 揭示之圖式及元件符號。 表單编號A0101 第5頁/共12頁 M413944 [0014] [0015] [0016] [0017] [0018] 100年08月02日修正替换頁 本創作利用聲音控制漏電流遮斷器之聲控裝置可增 設於目前切換式漏電流遮斷器控制裝置内，以在不改變 既有漏電流遮斷器的結構而執行利用聲音控制漏電流遮 斷器之動作。 請參閲第一、二圖，第一圖係說明本創作之利用聲 音控制漏電流遮斷器之裝置之方塊示意圖；第二圖係說 明利用本創作之以聲音控制漏電流遮斷器之裝置進行以 聲音控制漏電流遮斷器之語音辨識模組方塊示意圖。本 創作利用聲音控制漏電流遮斷器之聲控裝置包含有一語 音取音元件（1)、一語音辨識元件（2)及一微處理 器元件（3);其中： 語音取音元件（1 )，用於擷取一語音訊號，係裝 設在電源開關箱【圖中未表示】之外部空間，用來感測 觸發該聲控裝置； 語音辨識元件（2)，係與該語音取音元件（1 ) 連結，並對該語音取音元件（1 )所擷取之語音訊息進 行梅爾倒頻譜係數處理、隱藏式馬可夫模型處理、類神 經網路處理與動態時間校準處理，以對原始語音信號求 取有利於語音辨識之語音特徵參數，以作為語音訓練與 辨識之樣本； 微處理器元件（3)，係於接收到該語音辨識元件 (2)所輸出之語音辨識結果後，將一驅動訊號輸出至 該漏電流遮斷器（4)，以驅動對應的電源開關箱中之 漏電流遮斷器（4)進行切換動作。The voice data is sent to construct the voice feature parameters, and the voice recognition result is obtained by a special form number A0101, page 4/12 pages of the attributive language, and the material U. The processor control component, after receiving the identification result of the voice signal received by the identification component, drives a driving signal to the leakage current interrupter to activate the power switch box control circuit. [_] The material S of the towel material S, the remaining material identification mode uses Dynamic Time Warping technology architecture, for Speaker Dependent speech recognition, by voice recognition microcontroller (Speech Recognition Microcontroller), improve the performance of speech recognition. [0011] wherein the linguistic speaker recognition component is based on HMM/neural network hybrid for speaker independent speech recognition by speech recognition. Microcontroller (Speech Rec〇gni_ tion Microcontroller) 'Improve the performance of speech recognition capability [0012] According to the present invention, the sound control device for controlling the leakage current interrupter by using the sound is not changed by the existing leakage current interrupter operation. The option added in the mode to control the leakage current interrupter coexists. When the voice control is not triggered, the user can still use the original switching mode of the leakage current interrupter to control the leakage current interrupter. [Embodiment] [0013] For a more complete and clear disclosure of the technical means, the purpose of the creation, and the effect of the present invention, the following detailed description is provided, and the drawings and component symbols are also referred to. Form No. A0101 Page 5 / Total 12 pages M413944 [0014] [0017] [0018] Correction of the replacement page of this year, the creation of a voice control device using a sound control leakage current interrupter It is added to the current switching leakage current interrupter control device to perform the operation of controlling the leakage current interrupter by sound without changing the structure of the existing leakage current interrupter. Please refer to the first and second figures. The first figure shows the block diagram of the device for using the sound control leakage current interrupter. The second figure shows the device for controlling the leakage current interrupter by using the present invention. A block diagram of a speech recognition module for controlling a leakage current interrupter by sound. The voice control device using the sound control leakage current interrupter comprises a voice sounding component (1), a voice recognition component (2) and a microprocessor component (3); wherein: the voice sounding component (1), For capturing a voice signal, it is installed in an external space of a power switch box (not shown) for sensing and triggering the voice control device; a voice recognition component (2), and the voice sound pickup component (1) Linking and performing the Mel cepstral coefficient processing, the hidden Markov model processing, the neural network processing and the dynamic time calibration processing on the voice information captured by the voice pickup component (1) to obtain the original voice signal Taking the speech feature parameters favorable for speech recognition as a sample of speech training and recognition; the microprocessor component (3) is to receive a driving signal after receiving the speech recognition result output by the speech recognition component (2) The leakage current interrupter (4) is output to drive the leakage current interrupter (4) in the corresponding power switch box to perform a switching operation.
表單編號A0101 第6頁/共12頁 M413944 100年08月02日修正替換頁 .[0019] 依據本創作之利用聲音控制漏電流遮斷器之聲控裝 置，其中，該語音辨識元件（2)包含： [0020] 一語音音效卡（2 1 )，係將該語音取音元件（1 )所接收的語音訊號由類比訊號轉換至數位信號，以進 行系統之數位化處理，並進一步擷取該使用者所發出的 語音資料，以建構出語音特徵參數； [0021] » • • 一特定語者語音辨識元件（22)，需使用者事先 語音訓練，係依據語音音效卡（2 1 )所建構出之語音 特徵參數與事先儲存於語音辨識元件（2)内之特定語 者語音基於動態時間校準（Dynamic time warping) 字詞比對而得，係藉由語音辨識微控制器（Speech Recognition Microcontroller) 所建構； 以及 [0022] 一非特定語者語音辨識元件（2 3)，不需使用者 事先語音訓練，係依據語音音效卡（2 1 )所建構出之 語音特徵參數透過隱藏式馬可夫模型（Hidden Markov Model)處理、類神經網路（Neural Network)演算法語 音組合所成的字音轉換（Text-to-Speaker Inde-pendent)而得，係藉由語音辨識微控制器（Speech Recognition Microcontroller)所建構。 [0023] 據此，當緊急危難事件發生時，位在空間内的特定 語者或非特定語者均能透過語音的方式在最短的時間内 控制漏電流遮斷器進行切換動作。 [0024] 依據本創作之利用聲音控制漏電流遮斷器之聲控裝 置，由於其係在不改變現有漏電流遮斷器操作模式下所 表單編號A0101 第7頁/共12頁 [0025] 100年08月02日梭正替换頁 [0026] 增設的選項控制裝置，故可與原有藉切換方式控制漏電 流遮斷器之方法並存，於該裝置未被觸發時，使用者仍 可利用漏電流遮斷器原有的切換方式進行控制。 本創作已藉上述較佳實施例加以說明，為以上所述 者，僅為本創作之較佳實施例，並非用來限定本創作之 實施範圍。凡依本創作申請專利範圍所述之技術特徵及 精神所為之均等變化與修飾，均應包含於本創作之申請 專利範圍内。 綜上所述，本創作實施例確能達到所預期之使用功 效，又其所揭露之具體構造，不僅未曾見諸於同類產品 中，亦未曾公開於申請前，誠已完全符合專利法之規定 與要求，爰依法提出新型專利之申請，懇請惠予審查， 並賜准專利，則實感德便。 【圖式簡單說明】 第一圖：係說明本創作之利用聲音控制漏電流遮斷器之 裝置之方塊示意圖 第二圖：係說明利用本創作之以聲音控制漏電流遮斷器 之裝置進行以聲音控制漏電流遮斷器之語音辨識模組方 塊示意圖 【主要元件符號說明】 (1) 語音取音元件 (2) 語音辨識元件 (2 1 )語音音效卡Form No. A0101 Page 6 of 12 M413944 Correction Replacement Page, August 02, 100. [0019] According to the present invention, a voice control device for controlling a leakage current interrupter using a sound, wherein the voice recognition component (2) includes [0020] A voice sound card (2 1 ) converts the voice signal received by the voice sounding component (1) from an analog signal to a digital signal to perform digital processing of the system, and further utilizes the use. The voice data sent by the person to construct the voice feature parameters; [0021] » • • A specific speaker voice recognition component (22) requires the user's prior voice training, which is constructed according to the voice sound card (2 1 ) The speech feature parameters are compared with the speech of the specific speaker previously stored in the speech recognition component (2) based on the dynamic time warping word, and are performed by a speech recognition microcontroller (Speech Recognition Microcontroller). Constructing; and [0022] a non-specific speaker speech recognition component (23), which does not require prior user speech training, is based on the speech constructed by the speech sound card (2 1 ) The parameters are obtained by Hidden Markov Model and the text-to-Speaker Inde-pendent of the Neural Network algorithm. Constructed by the Sensor (Speech Recognition Microcontroller). According to this, when an emergency distress event occurs, the specific speaker or the non-specific speaker located in the space can control the leakage current interrupter to perform the switching operation in the shortest time by means of voice. [0024] According to the present invention, the sound control device for controlling the leakage current interrupter by using the sound, because it is not changing the existing leakage current interrupter operation mode, the form number A0101 is 7 pages/12 pages [0025] 100 years On August 02, the shuttle is replacing the page [0026]. The option control device is added, so it can coexist with the original method of controlling the leakage current interrupter by switching mode. When the device is not triggered, the user can still use the leakage current. The original switching mode of the interrupter is controlled. The present invention has been described in terms of the preferred embodiments described above, and is merely a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Equivalent changes and modifications to the technical characteristics and spirit described in the scope of the patent application shall be included in the scope of the patent application. In summary, the present embodiment can achieve the expected use efficiency, and the specific structure disclosed therein has not been seen in similar products, nor has it been disclosed before the application, and has fully complied with the provisions of the Patent Law. And the request, the application for a new type of patent in accordance with the law, please give us a review, and grant a patent, it is really sensible. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a device for controlling a leakage current interrupter using the sound of the present invention. FIG. 2 is a view showing a device for controlling a leakage current interrupter by using the present invention. Block diagram of voice recognition module for sound control leakage current interrupter [Key component symbol description] (1) Voice pickup component (2) Voice recognition component (2 1) Voice sound card
[0027] [0028] [0029] [0030] [0031][0027] [0030] [0030]
表單編號A0101 第8頁/共12頁 M413944 [0032] ( 2 2 )特定語者語音辨識元件 [0033] ( 2 3 )非特定語者語音辨識元件 [0034] (3) 微處理器元件 [0035] (4) 漏電流遮斷器 表單編號A0101 100年08月02日修正替换頁 第9頁/共12頁Form No. A0101 Page 8 of 12 M413944 [0032] ( 2 2 ) Special Speaker Speech Recognition Element [0033] ( 2 3 ) Non-Specific Speaker Speech Recognition Element [0034] (3) Microprocessor Element [0035 ] (4) Leakage Current Interrupter Form No. A0101 Correction Replacement Page, Page 9/12