本文選題：新息　切入點(diǎn)：神經(jīng)網(wǎng)絡(luò)　出處：《桂林理工大學(xué)學(xué)報(bào)》2017年01期 　論文類型：期刊論文

【摘要】：實(shí)際應(yīng)用中卡爾曼濾波的動(dòng)態(tài)噪聲和觀測(cè)噪聲往往不是固定的,提出運(yùn)用新息自適應(yīng)卡爾曼濾波法則對(duì)噪聲進(jìn)行實(shí)時(shí)估計(jì)。同時(shí),為了克服新息自適應(yīng)卡爾曼濾波要求系統(tǒng)模型過分精確的缺點(diǎn),基于神經(jīng)網(wǎng)絡(luò)的新息自適應(yīng)卡爾曼濾波算法,通過使用神經(jīng)網(wǎng)絡(luò)進(jìn)行誤差補(bǔ)償來提高濾波性能,并運(yùn)用于高速公路變形監(jiān)測(cè)中。結(jié)果表明:該模型與原始觀測(cè)值的殘差全部降到了0.25 mm以下,說明該模型精度很高,在實(shí)際應(yīng)用中具有一定的可行性。
[Abstract]:The dynamic noise and observation noise of Kalman filter are often not fixed in practical application. A novel adaptive Kalman filter is proposed to estimate the noise in real time. In order to overcome the shortcoming that the adaptive Kalman filter of innovation requires the system model to be too accurate, the adaptive Kalman filter algorithm based on neural network is used to improve the filtering performance by using neural network to compensate for the error. The results show that the residual difference between the model and the original observation value has been reduced to less than 0.25 mm, which indicates that the model has high precision and is feasible in practical application.
【作者單位】： 桂林理工大學(xué)廣西空間信息與測(cè)繪重點(diǎn)實(shí)驗(yàn)室;桂林理工大學(xué)測(cè)繪地理信息學(xué)院;桂林理工大學(xué)廣西礦冶與環(huán)境科學(xué)實(shí)驗(yàn)中心;
【基金】：國(guó)家自然科學(xué)基金項(xiàng)目(41461089) 廣西自然科學(xué)基金項(xiàng)目(2014GXNSFAA118288) 廣西“八桂學(xué)者”崗位專項(xiàng)經(jīng)費(fèi)項(xiàng)目 廣西空間信息與測(cè)繪重點(diǎn)實(shí)驗(yàn)室項(xiàng)目(桂科能140452402;130511402) 廣西礦冶與環(huán)境科學(xué)實(shí)驗(yàn)中心項(xiàng)目(KH2012ZD004) 廣西研究生教育創(chuàng)新計(jì)劃項(xiàng)目(YCSZ2014151;YCSZ2012083)
【分類號(hào)】：U412.24
，

本文編號(hào)：1639256