基于深度學(xué)習(xí)的短時(shí)交通流預(yù)測(cè)
本文選題:短時(shí)交通流預(yù)測(cè) 切入點(diǎn):深度學(xué)習(xí) 出處:《青島大學(xué)》2017年碩士論文 論文類型:學(xué)位論文
【摘要】:隨著城市智能交通系統(tǒng)的不斷完善,城市道路上安裝了大量的檢測(cè)器設(shè)備來(lái)采集道路和車(chē)輛的信息數(shù)據(jù),通過(guò)分析檢測(cè)器檢測(cè)到的大量數(shù)據(jù),為從數(shù)據(jù)中發(fā)現(xiàn)交通規(guī)律,支持城市交通優(yōu)化,短時(shí)交通流分析是當(dāng)前主要研究問(wèn)題之一。卷積神經(jīng)網(wǎng)絡(luò)避免了從數(shù)據(jù)中人工提取特征的問(wèn)題,采用卷積神經(jīng)網(wǎng)絡(luò)科自動(dòng)提取交通流量的時(shí)空特征,具有較好的應(yīng)用價(jià)值。基于此本文采用卷積神經(jīng)網(wǎng)絡(luò)的方法對(duì)短時(shí)交通流進(jìn)行研究預(yù)測(cè),主要內(nèi)容如下:本文通過(guò)構(gòu)建卷積神經(jīng)網(wǎng)絡(luò)模型,將路網(wǎng)的交通流數(shù)據(jù)轉(zhuǎn)化為交通流擁堵級(jí)別數(shù)據(jù),實(shí)現(xiàn)基于此類數(shù)據(jù)的預(yù)測(cè)時(shí)間小于5分鐘的交通狀況預(yù)測(cè),面對(duì)特定類型的樣本較小的情況,利用遷移學(xué)習(xí)的思路增加訓(xùn)練集的數(shù)據(jù)量和提升模型的預(yù)測(cè)性能。本文的主要工作如下:(1)針對(duì)短時(shí)交通流預(yù)測(cè)問(wèn)題,提出了采用卷積神經(jīng)網(wǎng)絡(luò)解決,分析了方案中卷積神經(jīng)網(wǎng)絡(luò)中卷積核大小對(duì)模型預(yù)測(cè)準(zhǔn)確度的影響,采用秦皇島連續(xù)并且缺失值最少的15天的數(shù)據(jù)集,實(shí)現(xiàn)了1分鐘和5分鐘的流量預(yù)測(cè)。并實(shí)現(xiàn)了高峰期和全天時(shí)段數(shù)據(jù)的預(yù)測(cè)。(2)針對(duì)數(shù)據(jù)訓(xùn)練樣本過(guò)少,導(dǎo)致模型過(guò)擬合問(wèn)題,采用遷移學(xué)習(xí)的思想,通過(guò)不同時(shí)段數(shù)據(jù)集的不同以及在訓(xùn)練集中增加隨機(jī)擾動(dòng)的方法,增大了數(shù)據(jù)樣本,提高了模型預(yù)測(cè)準(zhǔn)確性。
[Abstract]:With the continuous improvement of urban intelligent transportation system, a large number of detectors are installed on urban roads to collect information data of roads and vehicles. In order to support urban traffic optimization, short-term traffic flow analysis is one of the main research problems. Convolution neural network avoids the problem of manually extracting features from the data, and uses convolution neural network to automatically extract the space-time characteristics of traffic flow. Based on this, the method of convolution neural network is used to study and predict the short-term traffic flow. The main contents are as follows: this paper constructs the model of convolution neural network. The traffic flow data of the road network are transformed into traffic congestion level data, and the traffic condition prediction with less than 5 minutes prediction time based on this kind of data is realized. In the case of small sample size of specific type, the traffic flow data of the road network is transformed into the traffic congestion level data. The main work of this paper is as follows: (1) to solve the problem of short-term traffic flow prediction, a convolution neural network is proposed to solve the problem. The effect of convolution kernel size on prediction accuracy in convolution neural network is analyzed. The data set of Qinhuangdao continuous with the least missing value for 15 days is used. 1 minute and 5 minute traffic forecasting is realized. The prediction of peak period and whole day data is realized. In view of too few data training samples, the model is overfitted and the idea of transfer learning is adopted. Through the different data sets in different periods and the method of adding random disturbance in the training set, the data samples are enlarged and the prediction accuracy of the model is improved.
【學(xué)位授予單位】:青島大學(xué)
【學(xué)位級(jí)別】:碩士
【學(xué)位授予年份】:2017
【分類號(hào)】:U491.14;TP18
【參考文獻(xiàn)】
相關(guān)期刊論文 前10條
1 田晶;楊玉珍;陳陽(yáng)舟;;短時(shí)交通流量?jī)煞N預(yù)測(cè)方法的研究[J];公路交通科技;2006年04期
2 翁小雄;譚國(guó)賢;姚樹(shù)申;黃征;;城市交叉口交通流特征與短時(shí)預(yù)測(cè)模型[J];交通運(yùn)輸工程學(xué)報(bào);2006年01期
3 張赫,王煒,顧懷中;聚類分析和逐步回歸法在車(chē)道流量預(yù)測(cè)中的綜合應(yīng)用(英文)[J];Journal of Southeast University(English Edition);2005年03期
4 徐今強(qiáng),劉智勇;交通流的時(shí)間序列建模及預(yù)測(cè)[J];五邑大學(xué)學(xué)報(bào)(自然科學(xué)版);2004年03期
5 韓超,宋蘇,王成紅;基于ARIMA模型的短時(shí)交通流實(shí)時(shí)自適應(yīng)預(yù)測(cè)[J];系統(tǒng)仿真學(xué)報(bào);2004年07期
6 唐明,陳寶星,柳伍生;基于相空間重構(gòu)的短時(shí)交通流分形研究[J];山東交通學(xué)院學(xué)報(bào);2004年01期
7 黃潂,陳森發(fā),周振國(guó),亓霞;城市交通流量的非線性混沌預(yù)測(cè)模型研究(英文)[J];Journal of Southeast University(English Edition);2003年04期
8 宗春光,宋靖雁,任江濤,胡堅(jiān)明;基于相空間重構(gòu)的短時(shí)交通流預(yù)測(cè)研究[J];公路交通科技;2003年04期
9 達(dá)慶東,段里仁;交通流非參數(shù)回歸模型[J];數(shù)理統(tǒng)計(jì)與管理;2003年04期
10 宮曉燕,湯淑明;基于非參數(shù)回歸的短時(shí)交通流量預(yù)測(cè)與事件檢測(cè)綜合算法[J];中國(guó)公路學(xué)報(bào);2003年01期
相關(guān)碩士學(xué)位論文 前1條
1 陳綱;基于灰色理論和BP神經(jīng)網(wǎng)絡(luò)交通流預(yù)測(cè)模型研究[D];哈爾濱工業(yè)大學(xué);2006年
,本文編號(hào):1609577
本文鏈接:http://sikaile.net/kejilunwen/daoluqiaoliang/1609577.html