基於時空CNN-LSTM模型與多特徵資料分類器之道路交通風險預測暨個人化應用;Traﬀic Risk Prediction and Personal Application using CNN-LSTM Model and Classifier over Spatio-Temporal and Multi-Attribute Traﬀic Data

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Communication Engineering > Electronic Thesis & Dissertation > Item 987654321/89700

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/89700

Title:	基於時空CNN-LSTM模型與多特徵資料分類器之道路交通風險預測暨個人化應用;Traﬀic Risk Prediction and Personal Application using CNN-LSTM Model and Classifier over Spatio-Temporal and Multi-Attribute Traﬀic Data
Authors:	劉佩怡;Liu, Pei-Yi
Contributors:	通訊工程學系
Keywords:	時空交通資料;特徵嵌入;交通風險預測;Spatio-Temporal Data;Feature Embedding;Traffic Risk Prediction;CNN;LSTM
Date:	2022-08-30
Issue Date:	2022-10-04 11:53:15 (UTC+8)
Publisher:	國立中央大學
Abstract:	由於都市化而產生許多的交通問題造成人員傷亡以及經濟損失，於是交通風險預測與防範已成為重要的議題，隨著物聯網和人工智慧的發展，智慧運輸系統已成為現今改善交通的趨勢，人們不但可以利用無線通訊設備即時的溝通以取得最新的路況資訊，也能透過大數據分析預測未來的交通情況。然而多數研究已利用交通時空資料來實現交通風險預測，但缺乏考慮時空複雜性，導致無法有效提取交通狀態特徵。有鑑於此，本研究首先利用皮爾森相關係數探討並證明交通事故與時間和空間具有高度關聯性，再基於 CNN-LSTM 模型採用多特徵之交通時空資料進行道路交通風險預測，其中考慮了歷史事故、天氣、交通狀態和時間等特徵，並結合特徵嵌入方法進行 CNN-LSTM 道路交通風險預測模型之設計與訓練，使得該模型適用於實際交通環境。在實驗模擬上，本研究針對「單一交通事故資料」與「多特徵交通時空資料」等兩面向與不同預測模型進行效能分析比較，實驗結果表明，由於本研究方法基於 CNN-LSTM 模型結合有效提取外部因素的特徵嵌入與 Mask 層，於是 CNN-LSTM 道路交通風險預測模型可快速收斂且有較低的損失值，此模型具有更好的道路交通風險預測成效;最後，再進一步探討不同事故因素與交通事故嚴重程度的關係，本研究比較三種分類器:決策樹、隨機森林和 XGBoost，依用路人族群針對個人交通事故資料進行分類與分析，從中探討用路人對於交通事故因素的相關性。;With urbanization, many traﬀic problems have resulted in casualties and economic losses. Therefore, traﬀic risk prediction and prevention have become important issues. Due to the development of the Internet of Things(IoT)and Artificial Intelligence (AI), Intelligent Transportation System(ITS)have become the trend for improving transporta- tion. People can not only use wireless communication devices to communicate in real-time to obtain the latest road condition information but also predict future traﬀic conditions through big data analysis. However, many studies have used traﬀic spatio-temporal data to achieve traﬀic ac- cident prediction, but have not considered the spatio-temporal complexity, which leads to the inability to effectively extract traﬀic features. Given this, this study first uses the Pearson correlation coeﬀicient to explore and prove that traﬀic accidents are highly cor- related with time and space. Next, we use multi-attribute traﬀic spatio-temporal data to predict traﬀic risks based on the CNN-LSTM model, which considers historical accidents, weather, traﬀic condition, time, and so on. And then we combined the feature embedding method to design and train the “CNN-LSTM Road Traﬀic Risk Prediction Model”, making the model suitable for the actual traﬀic environment. In terms of experimental simulation, this study analyzes and compares the performance of “single traﬀic accident data”and“multi-attribute traﬀic spatio-temporal data ”with different prediction models. The results show that the CNN-LSTM Road Traﬀic Risk Prediction Model can converge quickly and effectively with a lower loss value, Our model has better road traﬀic risk prediction performance; Finally, to further explore the relationship between different ac- cident factors and the severity of traﬀic accidents. This study compares three classifiers: decision tree, random forest, and XGBoost to classify and analyze the personal traﬀic accident data of drivers, and then explore the importance of traﬀic accident factors.
Appears in Collections:	[Graduate Institute of Communication Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	96	View/Open

社群 sharing

Loading...