RelSIM aims to develop a practical, scalable, and cost-efficient solution for voice user interfaces (VUIs) in industrial environments. At its core, the project combines modern AI technologies—particularly Large Language Models (LLMs) and Reinforcement Learning (RL)—to systematically support and improve development, validation, and optimization processes.
A key focus is enabling these voice interfaces to run on low-power hardware without relying on cloud connectivity. This ensures both real-time responsiveness and high data security. At the same time, the project addresses the transferability of VUIs across different machines and systems, allowing for broad industrial adoption.
This approach is especially relevant for small and medium-sized enterprises (SMEs), which benefit from reduced complexity, lower costs, and faster integration into existing infrastructures.
RelSIM delivers tangible value for industrialcompanies across multiple dimensions:
Overall, RelSIM strengthens the innovation capacity and global competitiveness of SMEs, particularly in the machinery and manufacturing sectors.
RelSIM addresses key challenges in the industrial adoption of voice control: high development costs, limited scalability, and reliance on cloud infrastructures. Building on insights from the predecessor project DialoKIM, RelSIM advances a more robust and scalable approach.
The solution combines advanced AI technologies such as Large Language Models (LLMs), Reinforcement Learning (RL), and Causal AI with Digital Twins and Human-in-the-Loop concepts. The goal is to significantly streamline and partially automate the development, validation, and optimization of voice interfaces—while ensuring that human experts remain actively involved to maintain quality, safety, and practical relevance.
Digital Twins enable realistic, virtual testing of systems at early development stages, reducing risk and accelerating iteration cycles. At the same time, RL allows continuous improvement of interactions based on both simulated and real user behavior.
The result is a flexible and adaptive system that can efficiently develop and refine voice interfaces for a wide range of industrial applications—laying the foundation for a new generation of intelligent, scalable human-machine interaction.
SABO Mobile IT GmbH in Bühl produces software on behalf of German industrial customers with a focus on HMI (Human Machine Interfaces) and solutions in the sense of "Industry 4.0", both for planning and supporting industrial production processes (e.g. at Daimler, Audi and VW), as well as for supplementing devices and machines in the food technology, medical technology and mechanical engineering sectors.
More information
The core activity of EES Beratungsgesellschaft mbH is IT-related business process optimization in industry and SMEs. EES is a team of experts from various technical disciplines and advises companies and engineers on topics relating to visualization and simulation in the areas of development, sales and virtual commissioning. With its own open source solution, EES enables 3D simulation and virtual prototyping of systems to accelerate and secure their development and construction.
In project acquisition, planned systems can be visualized in dialogue with the customer for effective sales. EES has proven its know-how and expertise in the field of Industry 4.0 through numerous projects for well-known companies.
More information
The start-up agitum GmbH is a spin-off from the Karlsruhe Institute of Technology (KIT) and was founded in April 2021. The founders and employees are experts and researchers in the field of machine learning with a focus on natural language processing, as well as in the areas of process and requirements engineering, process automation and digitalization.
As a KIT spin-off, agitum has experience in national and international industry and research projects. agitum currently uses ML methods in its customer projects to extract unstructured data from documents, among other things. For example, transformer-based methods of optical character recognition are used. agitum contributes to the project in particular in the automated data acquisition and the development of application-specific algorithms.
The Chair of Production Engineering and Industrial Organization (FBK) is part of the Department of Mechanical and Process Engineering at the Rhineland-Palatinate Technical University in Kaiserlautern. The focus of basic and application-oriented research is on machining and digital technologies for production systems.
There is extensive expertise in the design, construction and operation of machines in the production environment. Another core research topic is machine learning, which is being researched as part of several collaborative and basic research projects relating to industrial production. Important preliminary work also exists in the area of designing interactive user interfaces for machines in production, for example through augmented reality.
More informationGet in touch with our contacts now to find out more about the innovative DialoKIM research project and the benefits for your personal use case.
Master of Computer Science
Managing Director SABO Mobile IT GmbH
Engelstr. 6
77815 Bühl
Germany
Phone:
+49 7227 4946
Managing Director
EES Beartungsgesellschaft mbH
Adalbert-Stifter-Str. 8
76275 Ettlingen
Germany
Research Assistant
Chair of Manufacturing Engineering
and Industrial Organization
Gottlieb-Daimler-Str.
67663 Kaiserslautern
Germany
Managing Director Technology & Founders
agitum GmbH
Schönbornstraße 11
76661 Philippsburg
Germany