KI_IDENT: Machbarkeitsstudie zur KI-gestützten Identifikation von neuralgischen Punkten in Verkehrsnetzen basierend auf Verkehrsflussdaten von Versorgungsverkehren
Partner: SETLabs Research GmbH, München (Tochter der Virtual Vehicle Research GmbH)
Project Duration: 01/2023 - 06/2024
Funding: mFUND, Bundesministerium für Digitales und Verkehr
Problem Definition:
Supply transports of critical and social infrastructure (e.g. emergency services, fire department, bus transport, waste disposal companies) have to reach their destinations quickly, safely and reliably. In practice, this is made difficult by recurring obstructions, such as parking violations, physical blockades and infrastructural deficiencies. The detection of such neuralgic points in the road infrastructure supply network is a major interest of planning authorities and the above-mentioned operators.
Project Objective:
The project objective is to conduct a feasibility study for a novel software-based information system that identifies neuralgic traffic points. A harmonized data pool of traffic flow data from different public vehicle fleets will be used to train AI models for automated detection of obstructions. In this process, data analyses and result processing are carried out along the requirements of operators and planning authorities elaborated in usecases.
Implementation:
The feasibility study is performed in several phases: First, relevant data pools are identified and harmonized in a common data pool. If necessary, additional movement data will be collected by technically equipping vehicles that do not yet have the corresponding equipment. On this data basis, AI models are trained for the detection of neuralgic traffic points. On the basis of concrete use cases, which are developed with the stakeholders involved, information and decision-making aids relevant to planning and operations can be designed. For this purpose, target group-specific processing and visualization will be carried out in an interactive, software-supported information system.