ML4RDE - Machine lerarning for real driving emissions

Project: Research project

Project Details

Description

Objectives and Motivation: • The estimation of energy consumption of a driver moving along some road currently is based on statistical methods that give good but not very precise forecasts. Within this Project, these forecasts shall be made much more precise without losing explainability or reliability. Methodology: • Based on real-driver data and combinations of automata-learning and machine learning methods, more accurate models of driver behaviour shall be constructed that give more precise forecasts. For comparison purposes a purely deep-learning approach also is followed. Expected Results: • More precise driver models that are able to predict the energy usage over arbitrary roads and given a certain driving “style” more precisely, while still retaining modest computational overhead and being “real-time” (i.e., no GPUs required, couple of minutes for the computation).
StatusFinished
Effective start/end date1/02/2331/01/24

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.