Applied Research

Topic: Efficient and characteristic data generation

Project partners: Continental Automotive Technologies GmbH, Mercedes-Benz AG, Valeo Schalter und Sensoren GmbH, AVL Software and Functions GmbH, AVL Deutschland, FZI Research Center for Information Technology, Luxoft GmbH, Deggendorf institute of technology

The core idea of the just better DATA project is the development and implementation of AI-based methods and tools to collect data efficiently and with high accuracy and process it at the edge instead of generating unnecessarily large amounts of data.
During the three-year project, a process will be developed and implemented as a prototype that will provide "better data" for training AI models for autonomous driving in the future.

• Compilation of the requirements for the new system
• Development of an edge platform for recording, pre-processing and selecting data in the vehicle in real time
• Optimization of the classic data loop through the use of smart algorithms
• Preparation of vehicle equipment for data acquisition and recording of real data to generate a data set that enables industrial use for all partners

Duration: 07/2023 - 06/2026

Link to the research project

Topic: Embedded Excellence - Driving Dynamics with AI

Project partners: Infineon Technologies AG, ZF Friedrichshafen AG, samoconsult GmbH, RWTH Aachen, University of Lübeck

Aim: Cloud-independent, real-time embedded AI vehicle dynamics system that extends the classic control of the actuators in the vehicle with the advantages of AI. Future autonomous driving and assistance systems will be based on this fundamental functionality, which is deliberately open to OEMs.

• Development of suitable edge machine learning methods for (incremental) learning
• Development of an offline (batch learning) and online (incremental learning) training environment
• Virtualisation of hardware components
• Fusion of the real and simulated data
• Identification of anomalies within the communication in the Edge AI subsystem of an ECU network
• Implement and evaluate cloud-independent, real-time embedded AI vehicle dynamics system
   

Duration: 09/2021–08/2024

Link to the research project

Topic: In-Vehicle Big Data Analytics System for the development of data-based algorithms by using intelligent edge nodes

Project partner: Deggendorf Institute of Technology (DIT)

The INSTATE project uses intelligent edge devices to generate metadata in the vehicle. These are connected to a vehicle's sensors. The metadata generated can be used by a Big Data platform to understand the current scenario. In addition, data, including metadata, is only recorded as required.

• Data-driven algorithms for autonomous driving using intelligent edge nodes
• Creation of meta data directly in the vehicle
• Converting the sensor data into compressed streams
• Backend connection via 5G gateway
   

Duration: 07/2021–12/2024

Topic: 5G in the National Park Region

Project partners: Gemeinde Spiegelau, THD, DB Regio AG, IQ MEDWORKS GmbH, rBITech GmbH, Elektra Solar GmbH, 3D Reality Maps GmbH, Spekter GmbH

The project "5G in the National Park Region" deals with the application of 5G for various deployment scenarios in the agricultural and at the same time very forested low mountain region around the Bavarian Forest National Park with communities characterised by tourism. The project concept provides for the use of a 5G infrastructure - consisting of mobile and flying 5G relay stations as well as campus networks applications in the four fields of action tourism (mobility and transport), agriculture, forestry, rescue services and smart municipal infrastructures.

• Connecting the sensors and measurement technology of an autonomous bus via 5G to the cloud development of an offline (batch learning) and online (incremental learning) training environment
• The autonomous bus is extended by an intelligent data logger controlled via 5G.
• Algorithms for evaluating the data
   

Duration: 01/2022–12/2024

Link to the research project