University of Thessaly (UTH)
Μεθοδολογικό πλαίσιο αναγνώρισης και ταυτοποίησης της δυναμικής της οδικής αστικής κυκλοφορίας
Abstract
dc:descriptionThe ultimate objective of the current thesis is the identification and understanding of the underlying spatiotemporal dynamics within an urban road traffic system through deepening in the potentials unlocked by correlation and causality analysis. Case study for the undertaken analysis is the city of Thessaloniki, a modern Greek urban area of medium size, while floating taxi (GPS data) and Bluetooth detections based travel time timeseries constitute the basic data input. Complementary travel time data coming from a field measurement campaign with 3 floating experimental vehicles were used (covering the assumption of general traffic). All data used originated from efforts of CERTH/HIT (Centre for Research and Technology Hellas / Hellenic Institute of Transport) to develop and maintain the ‘Thessaloniki Smart Mobility Living Lab’ serving research and innovation objectives. The basic research questions posed refer to i) the reliability of traffic data collected via the various sources regarding the general traffic and the potential complementarity among them, ii) the spatiotemporal relations among road paths and iii) the exploitation of knowledge derived from the timeseries based prediction models. In line with the above research questions, an integrated methodological framework was developed making use of the following tools respectively: i) classical statistical descriptive analysis, linear, logarithmic and generalized linear models, ii) univariate (descriptive features, sequence plots, linear and non linear measures calculation as autocorrelation and mutual information) & multivariate timeseries analysis (clustering, correlation, cross-correlation, Granger causality) and iii) timeseries based models development (SARIMAX, multiple linear regression, temporal causal modelling). The analysis validates a good reliability level of the provided information from both floating taxi data and Bluetooth detections in the examined road paths while it concludes on a proposal for their complementarity. The analysis provides also useful insight for the complex causal relations among the different road paths of the network validating the existing traffic intuition. Furthermore, the visualizations deriving from the advanced time series analysis presented (Temporal Causal Models) facilitate the understanding of the relations. The flexibility of the proposed methods in including additional time series / new ‘predictors’ (e.g. delay related time series) and the potential combination of time series analysis with machine learning techniques can probably help unlocking potentials for identifying tailored equations describing traffic characteristics at urban areas (i.e. tailored volume - delay functions).
Degree
thesis:*- Grantor dc:publisher
- University of Thessaly (UTH)
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Myrovali, Glykeria
- Μυρόβαλη, Γλυκερία
Subjects
dc:subject × 16- Οδικές μεταφορές
- Χρονοσειρές, Ανάλυση
- Αιτιότητα χρονοσειρών
- Δυναμικά συστήματα
- Χωροχρονικά φαινόμενα
- Αστική κυκλοφοριακή συμφόρηση
- Urban transport
- Timeseries analysis
- Granger causality
- Dynamic systems
- Spatiotemporal phenomena
- Urban road network management
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Πολιτικού Μηχανικού
- Engineering and Technology
- Civil Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/47942