To path indicator Subpages of “Newsdetail“ To navigation To quick access To footer with other services
To content
Faculty of mechanical engineering
Studierende sitzt am Arbeitsplatz und schaut auf den Laptop. © Aliona Kardash​/​TU Dortmund

LLM-supported decision support for AGVs

-
in
  • News
  • Ausschreibung wissenschaftlicher Arbeiten
  • Masterarbeit
The aim of this work is to develop an LLM-supported system for decision support in the selection of AGVs

The use of autonomous mobile robots (AMRs) and automated guided vehicles (AGVs) is steadily increasing in modern industry and logistics. However, due to the large number of available models and the different technical specifications, it is often difficult to select a suitable AGV for specific applications. Traditional decision-making processes are based on manual research and expert knowledge, which can be time-consuming and error-prone. The use of Large Language Models (LLMs) offers a promising opportunity to automatically extract technical data and compare it with individual requirements.

Objective

The aim of this work is to develop an LLM-supported system for decision support in the selection of AGVs. Automated web-scraping techniques are used to extract and structure technical data from manufacturer websites and evaluate it using an LLM. The system should enable users to define requirements for AGVs, which are then compared with the extracted data in order to make well-founded purchase recommendations. Finally, the developed methodology will be validated by comparing it with expert assessments.

Project plan

  1. Determination of AGV manufacturers
    • Identification of relevant manufacturers of autonomous mobile robots (AMRs/AGVs).
    • Selection of those with publicly available technical data (websites, PDFs, product catalogs).
  2. Definition of the parameters to be extracted
    • Definition of the most important technical specifications (e.g. payload, battery life, navigation technology, speed).
    • Classification into mandatory and optional parameters for better comparability.
  3. Web scraping: collection of AGV data
    • Automated retrieval of manufacturer pages, data sheets and documentation (HTML, PDFs, TXT).
    • Storage of raw data in a structured database for further processing.
  4. Extraction of the required parameters using LLM
    • Processing of the collected raw data with an LLM for targeted extraction of the relevant technical features.
    • Conversion of unstructured text data into a standardized, tabular form.
  5. LLM matching of extracted data with use case requirements
    • User specifies requirements (e.g. warehouse logistics, load capacity, environment).
    • LLM compares requirements with the extracted AGV data and makes a recommendation based on similarity and weighting.
  6. Validation: Comparison of LLM matching with expert matching
    • Comparison of the LLM-based recommendations with expert assessments.
    • Optimization of the matching logic through feedback and adjustment of weighting factors.

This video shows basically how steps 3 and 4 can work:

https://www.youtube.com/watch?v=Osl4NgAXvRk

 

Contact:

To top of page