提出了基于智能嗅覺(jué)系統(tǒng)的識(shí)別混合有毒氣體組分濃度的方法。該系統(tǒng)包括兩大部分: 有毒氣體傳感器陣列模塊和徑向基函數(shù)神經(jīng)網(wǎng)絡(luò)模塊。前者用于獲取反映有毒氣體組分的電信號(hào), 后者用于提高識(shí)別混合有毒氣體組分的選擇性, 降低氣體傳感器陣列中各個(gè)敏感器件的交叉靈敏度。徑向基函數(shù)神經(jīng)網(wǎng)絡(luò)具有很強(qiáng)的非線性并行處理能力和容錯(cuò)
能力, 實(shí)例分析取得了滿意的測(cè)量結(jié)果。
關(guān)鍵詞: 智能嗅覺(jué)系統(tǒng); 徑向基函數(shù)網(wǎng)絡(luò); 氣體傳感器陣列; 選擇性
Abstract: Intelligent senso r array system is suggested to detect low concent rat ion of m ixed gases1 It is compo sed of two block s, that is gas senso r array and radial basis funct ion neural netwo rk (RBFNN ) 1 The fo rmer is used to acquire re2 sponse signal, and the later is used to imp rove the select ivity of the gas senso r1 RBFNN hasmo re advantage than BPNN ,
such as fault to lerance, nonlinear mapp ing etc, thus, it can overcome the cro ss2sensit ivity among heterogeneous gas sen2 so r and get mo re sat isfied measurement result1
Key words: intelligent o lfact ion system; radial basis funct ion netwo rk; gas senso r array; select ivity