Title

PEAK-AWARE TRAFFIC PREDICTION WITH DEEP LEARNING MODELS

Abstract

Abstract

Accurately predicting traffic is crucial due to its impact on urban life in many aspects. Several statistical methods, machine learning, and deep learning approaches are applied to different traffic datasets. In general, traffic follows a stable behavior except for the morning and evening peaks which span a small period in time. To consider a prediction model to be accurate, it must demonstrate successful results during peak hours. In this thesis, a novel distance to mean weighting technique is presented that can be applied to any deep learning model by introducing a minor change in the loss function. The method also makes it possible to evaluate the performance of the models during peak hours in traffic. Traffic prediction is frequently used in estimating travel times and energy consumption (battery or fuel). In this thesis, in order to estimate both travel time and consumption, a driver simulation model based on Probabilistic Hybrid Automata is developed. The simulator generates speed-time data with respect to the given traffic and driver characteristics. The result is then used to estimate the consumption using an electric vehicle simulator.

Supervisor(s)

Supervisor(s)

FATIH ACUN

Date and Location

Date and Location

2022-07-29 10:00:00

Category

Category

MSc_Thesis