Project name

KIsSME

KIsSME project at a glance: https://www.kissme-projekt.de/

The project “künstliche Intelligenz zur selektiven echtzeitnahen Aufnahme von Szenarien- und Manöverdaten bei der Erprobung hochautomatisierter Fahrzeuge” (KIsSME) aims to evaluate driving situations using algorithms and AI, and to compare the results. Its main purpose is to reduce the large amounts of data generated during the drive to save storage space, power, and processing efforts. The project involves collaborative work among several partner companies, with each focusing on different aspects in collaboration with one another.

The R&D team of RA Consulting has been working on the development of interfaces for evaluating driving situations in vehicles during operation as part of this project, which was funded by the “Bundesministerium für Wirtschaft und Klimaschutz” (BMWi) from 2021 to 2023.

// Project partners
 
// Technologies

Development of a system for reliable criticality assessment of driving situations

Vehicles of the future must be able to make reliable driving decisions in the shortest possible time. The core of the KIsSME system is an artificial intelligence whose algorithm is trained and optimized until the criticality value for evaluating the driving situation is optimized. The goal is to have an artificial intelligence in the vehicle that correctly evaluates every driving situation, including unexpected ones.

Vehicles of the future must be able to make reliable driving decisions in the shortest possible time. The core of the KIsSME system is an artificial intelligence whose algorithm is trained and optimized until the criticality value for evaluating the driving situation is optimized. The goal is to have an artificial intelligence in the vehicle that correctly evaluates every driving situation, including unexpected ones.

Overview of core technologies

  • Machine Learning (AI)

    Machine Learning wird genutzt, um Fahrsituationen automatisch zu bewerten und zu klassifizieren. Durch die Analyse von Daten können Algorithmen trainiert werden, um bestimmte Ereignisse wie Unfälle oder gefährliche Fahrmanöver zu erkennen.

  • Neural networks (AI)

    Neuronale Netze können trainiert werden, um komplexe Muster in Daten wie Fahrverhalten und-Umgebung zu erkennen und zu bewerten. Durch die Verwendung von zufälligen Eingabesignalen können neuronale Netze in der Lage sein, unerwartete Situationen auf der Straße zu erkennen und deren Relevanz zu klassifizieren.

  • Data Management

    Eine effektive Datenvoranalyse ist entscheidend für die Bewertung von Fahrsituationen und die Identifizierung von Mustern und Trends in den Daten.

    Um eine erfolgreiche Datenvoranalyse durchzuführen, müssen die Daten gut organisiert und strukturiert sein, um sicherzustellen, dass die benötigten Informationen leicht zugänglich sind.

  • Simulations

    Simulationen können genutzt werden, um KI-Systeme im Bereich der Bewertung von Fahrsituationen zu trainieren, ohne das Risiko von Unfällen oder Verletzungen zu erhöhen.

    Durch das Erstellen realistischer Simulationen können verschiedene Szenarien getestet werden, um zu sehen, wie das KI-System auf bestimmte Fahrsituationen reagieren würde.

// Case study

From top to bottom: Autonomous work tool (AVL), vehicle demonstrator People Mover (Bosch) and autonomous car (FZI), whose data was used as a prototype in the project.

In today’s world, there is no industry or sector that is not driven by data. Data enables companies to optimize their processes, improve their products and services and make informed business decisions. The automotive industry is no different. In recent years, the amount of data generated in modern vehicles has increased dramatically. Lidar, radar, GPS and camera technology generate enormous amounts of data that need to be filtered and analyzed (Software Defined Vehicle). RA Consulting’s aim in the KIsSME project is to show how companies can automatically reduce the problem of the amount of data generated by supporting artificial intelligence and filter and evaluate relevant data.

Schedule consisting of data logging, data recording and data reduction tool.

The processing and analysis of data in modern vehicles pose a significant challenge. One of the major hurdles is to filter the relevant data in real time, allowing for the storage of only high-quality data to minimize energy consumption and analysis efforts. Questions such as “When is the data recorded?”, “What data is recorded?”, and “What is currently happening?” can be answered by a complex AI through continuous training. Making the right assessment of data criticality and relevance is essential for reliably evaluating and potentially storing critical driving situations.

Our focus in developing a solution lies in integrating software into the hardware of vehicles (OnBoard). We have concentrated on developing systems that enable the further processing of data in vehicles in a highly compatible manner. Additionally, we have worked on standardizing data formats to enhance the interoperability of different systems.

If you have any further questions or need additional information, feel free to ask.

This has helped to improve the efficiency and accuracy of data filtering and analysis.

Another challenge is to reduce the amount of “data garbage” generated in modern vehicles. “Data garbage” refers to data that is irrelevant and only consumes unnecessary energy and resources. We are enhancing our recording technology so that only relevant data is captured. For this purpose, we have developed a criticality metric that enables embedded systems not only to assess driving situations correctly but also to decide which data should be recorded and which should not. This significantly reduces the amount of “data garbage” and improves the efficiency of data filtering and analysis. The relevant data is then transferred via OnBoard loggers to software on the laptop, desktop PC (DiagRA® X), or to a digital twin or the cloud.

Our approach is to develop a concept for the future of the automotive industry. The current trend is towards service-oriented products, consisting of multiple components, with their subcomponents functioning as services. These components could be integrated into a complete system using, for example, OSGi-like architectures. By using OSGi-like architectures, we could create a highly modular system that allows individual components to be updated or replaced independently, thereby improving the flexibility and scalability of the system.

If you have similar projects and need support in data processing and analysis, we are at your disposal. Feel free to contact us at any time to learn more about our solutions and expertise.

Do you have any questions about the project?

Overview of the project structure and participants of KIsSME.

// Used tools
 

Functionality used:
Embedded diagnostic component

Go to DiagRA® Embedded

Functionality used:
MDF 4 recording

Go to DiagRA® X/DiagRA® X Viewer Pro

Functionality used:
Read data and validate format

Go to OpenScenario API

Functionality used:
OnBoard test scripts via emotive OTF

Go to Emotive OTX-Component

// Our contribution

What do we contribute to the
research project?

Image: icons8.com

Conceptualization of the On-Board System

  • Specification of a Reference System
  • Design of Interfaces for Software Components
  • Design of the System Architecture

Provision of extensive expertise on ASAM standards

  • ASAM OpenSCENARIO® , ASAM OpenTEST®, ASAM OpenDRIVE®, ASAM iLinkRT® and ASAM MDF/CCP/ODX
  • Interaction with working groups at the various participating organizations
  • Evaluation of standards and tools

Image: icons8.com

Image: icons8.com

Provision of measurement values

  • DiagRA® Embedded as a software interface
  • Embedded MC Server (EMC Server) for data acquisition and processing
  • Transmission to a control center
  • Provision of measurement values from ROS topics
// Our contribution

What do we contribute to the
research project?

Image: icons8.com

Conceptualization of the On-Board System

  • Specification of a Reference System
  • Design of Interfaces for Software Components
  • Design of the System Architecture

Image: icons8.com

Provision of extensive expertise on ASAM standards

  • ASAM OpenSCENARIO® , ASAM OpenTEST®, ASAM OpenDRIVE®, ASAM iLinkRT® andASAM MDF/CCP/ODX
  • Interaction with working groups at the various participating organizations
  • Evaluation of standards and tools

Image: icons8.com

Provision of measurement values

  • DiagRA® Embedded as a software interface
  • Embedded MC Server (EMC Server) for data acquisition and processing
  • Transmission to a control center
  • Provision of measurement values from ROS topics
// Benefits

What advantages does the project offer for the automotive industry and beyond?

Effective data filtering and evaluation in the vehicle

Significant reduction in non-relevant data

Continuous training of the AI and thus optimization of the criticality assessment of driving scenarios

Building up and transferring know-how in the field of artificial intelligence

Development of new components for autonomous driving for integration into existing overall systems

Research projects provide insights and market observations for future developments in the field of vehicle technology

// Nutzen

Welche Vorteile bietet das Projekt für die Automobilindustrie und darüber hinaus?

Effective data filtering and evaluation in the vehicle

Significant reduction in non-relevant data

Continuous training of the AI and thus optimization of the criticality assessment of driving scenarios

Building up and transferring know-how in the field of artificial intelligence

Development of new components for autonomous driving for integration into existing overall systems

Research projects provide insights and market observations for future developments in the field of vehicle technology

// Sponsors and supporters

Do you have any questions about the project or our research?

// Contact us

Dr. Frank Hantschel
Phone: +49 7251 9819 574
E-Mail: f.hantschel[@]rac.de

Images used on this page: rac.de, unsplash.com, icons8.com

Do you have any questions about the project or our research?

// Contact us

Dr. Frank Hantschel
Phone: +49 7251 9819 574
E-Mail: f.hantschel[@]rac.de

Images used on this page: rac.de, unsplash.com, icons8.com

Link to: Research and development at RA
Link to: Research and development at RA
OUR TOP PRODUCTS

DiagRA® D

DiagRA® X

DiagRA® X Viewer Pro

Silver Scan-Tool™

DiagRA® S

DiagRA® LE

RA® ODX Viewer

IAV Macara

DiagRA® Flash Station

DiagRA® Embedded

SUPPORT

Contact & Support

RA Download Center

Drivers

TRAININGS

Product trainings

Customized trainings

COMPANY

About us

Newsroom

Career

Worldwide distribution

Memberships

Subscribe to receive
news and updates from
RA Consulting

Compliance    •    Terms and Conditions    •    Privacy Policy    •    Legal Notice

© RA Consulting GmbH, 2024  •  USt.-Ident.-Nr.: DE143081464  •  HRB: 231127 ASAMAETAElektromobilität Süde-West
RA Consulting GmbH