可穿戴聲學(xué)傳感器通過(guò)校準(zhǔn)喉部振動(dòng)并將其轉(zhuǎn)化為合成語(yǔ)音，為語(yǔ)言障礙者提供了有效的溝通解決方案。本文，香港理工大學(xué)Zhongqing SU、南方科技大學(xué)周利民教授等在《ACS Sens》期刊發(fā)表名為“An Aerosol Jet-Printed Wearable Graphene/Cellulose Nanocrystal Acoustic Sensor for Speech Recognition”的論文，研究開(kāi)發(fā)了一種新型壓阻式聲學(xué)傳感器，通過(guò)氣溶膠噴射打印技術(shù)，采用聚氨酯（PU）薄膜封裝石墨烯/纖維素納米晶體（CNCs）進(jìn)行增材制造。該傳感器具有高度生物相容性和柔韌性，能夠精確測(cè)量變化的聲音壓力水平（SPL）。

實(shí)驗(yàn)結(jié)果表明，通過(guò)調(diào)節(jié)石墨烯濃度可調(diào)控傳感器的聲學(xué)靈敏度。當(dāng)石墨烯濃度為20%時(shí)，傳感器展現(xiàn)出9.7×10?? dB?1的高靈敏度，工作范圍覆蓋30至90 dB，聲壓級(jí)變化最小分辨率達(dá)10 dB，且聲壓級(jí)與傳感器測(cè)得的電阻變化呈線性相關(guān)。當(dāng)作為可穿戴設(shè)備貼附于受試者喉部時(shí)，該傳感器能精準(zhǔn)捕捉音色、節(jié)奏等細(xì)微發(fā)聲特征。結(jié)合基于支持向量機(jī)（SVM）的機(jī)器學(xué)習(xí)算法，該設(shè)備在識(shí)別數(shù)字（0-9）時(shí)達(dá)到95.9%的高準(zhǔn)確率，助力語(yǔ)言障礙者實(shí)現(xiàn)數(shù)字化溝通。

圖文導(dǎo)讀

圖 1. (a) Fabrication flow of sensor ink; (b) schematic of the acoustic sensor fabricated using AJP; (c) schematic diagram of the wearable acoustic device preparation process; (d) schematic of the acoustic device; (e) device worn on the skin; (f) cross-sectional TEM image of the GC-20%G.

圖2. Schematic of the sensor conductive path at initial condition (a) and sound acoustic pressure-loaded environment (b).

圖3. Acoustic response of the sensor. (a) Response of sensors at varying sound pressure level; (b) schematic diagram of the change in tunneling path of sensors with lower and higher graphene concentrations under sound pressure; (c) frequency response of the GC-20%G sensor under a representative SPL of 80 dB with different frequencies (from 10 to 500 Hz); (d) real-time response signal of the sensor to the signal with gradually increasing sound pressure level; (e) relative resistance changes with the 1st and 100th cyclic sound pressure levels; (f) three acoustic waveforms from repeated playback of the same audio segment at 80 dB SPL: original loudspeaker signal, iPhone recording, and GC-20%G sensor output.

圖4. Acoustic device for voice detection. (a) Wearable acoustic device attached on the throat; (b) relative resistance change of the sensor when three different volunteers say, “Merry Christmas”; (c?h) relative resistance change of the sensor in response to three different sentences, compared with the microphone signal; (c, f) is the signal response of “The Hong Kong Polytechnic University”; (d, g) is the signal response of “Hope everything goes well with you”; and (e, h) is the signal response of “What a nice weather”.

圖5. Acoustic devices for digits 0–9 recognition. (a) Signal of digits from 0 to 9 captured by the sensor; (b) schematic flowchart of speech recognition; (c) classification confusion matrix of digit 0–9 prediction versus the test data set; (d) recognition of a sequence of digits “007”.

小結(jié)

本研究采用AJP工藝成功制備了新型壓阻式石墨烯/碳納米管（CNC）聲學(xué)傳感器，并采用聚氨酯薄膜封裝以增強(qiáng)柔韌性與貼合皮膚的能力。石墨烯濃度為20%（GC-20%G）的傳感器展現(xiàn)出9.7×10?? dB?1的最高靈敏度，以及10 dB的最小聲壓級(jí)變化分辨率，從而實(shí)現(xiàn)精準(zhǔn)聲學(xué)檢測(cè)。該傳感器在30至90分貝聲壓級(jí)范圍內(nèi)表現(xiàn)出低滯后特性與優(yōu)異可逆性，并具備卓越的重復(fù)測(cè)量穩(wěn)定性：局部最大響應(yīng)與最小響應(yīng)的標(biāo)準(zhǔn)差值僅分別為0.035和0.0349。該可穿戴聲學(xué)設(shè)備固定于受試者喉部，能可靠捕捉聲帶振動(dòng)，從而精確識(shí)別語(yǔ)音模式、音調(diào)變化及獨(dú)特音頻信號(hào)。基于機(jī)器學(xué)習(xí)訓(xùn)練的SVM模型能以95.9%的準(zhǔn)確率識(shí)別傳感器采集的語(yǔ)音數(shù)字信號(hào)，證明其解讀語(yǔ)音信號(hào)語(yǔ)義內(nèi)容的能力。這些成果彰顯了傳感器的檢測(cè)與識(shí)別潛力，為語(yǔ)障人士的通信技術(shù)應(yīng)用拓展了廣闊前景。

展望未來(lái)，通過(guò)更大規(guī)模、更豐富多樣的數(shù)據(jù)集訓(xùn)練模型，有望實(shí)現(xiàn)更復(fù)雜語(yǔ)音模式的識(shí)別——從數(shù)字?jǐn)U展至完整詞匯與短語(yǔ)，從而拓展其在人機(jī)交互、醫(yī)療健康及無(wú)障礙技術(shù)領(lǐng)域的應(yīng)用場(chǎng)景。此外，將運(yùn)用分子動(dòng)力學(xué)模擬深入探究微尺度隧道效應(yīng)機(jī)制，通過(guò)定量分析為新一代高性能聲學(xué)傳感器的理性設(shè)計(jì)提供指導(dǎo)。

審核編輯 黃宇