May C, Suchy L, Franke J, Reitelshöfer S (2024)
Publication Type: Conference contribution
Publication year: 2024
Pages Range: 159-166
Conference Proceedings Title: Langbeiträge ASIM SST
Autonomous Mobile Systems (AMS) offer significant advantages in industry and private sectors by adapting to diverse and dynamic environments. To train these systems, large amounts of data are required, typically obtained from simulated environments. However, the creation of these environments is often labor-intensive. Here, we propose a generative pipeline that provides a streamlined approach to virtual training and testing while allowing users to apply automated methods including generative AI. Our pipeline consists of four, partly iterative main steps. The pipeline spans from the creation of individual assets to the utilization of the simulated environments. The pipeline is then implemented for an exemplary scenario, utilizing multiple methods including generative AI. Furthermore, we propose a novel application of our pipeline to provide robots with the capabilities to “imagine” virtual experiences. The presented pipeline not only simplifies the process of generating simulated environments, but also resembles a scalable framework for developing increasingly complex AMS.
APA:
May, C., Suchy, L., Franke, J., & Reitelshöfer, S. (2024). Towards Imaginative Robots: A Generative Pipeline for Simulated Environments. In Langbeiträge ASIM SST (pp. 159-166). München, DE.
MLA:
May, Christopher, et al. "Towards Imaginative Robots: A Generative Pipeline for Simulated Environments." Proceedings of the 27. Symposium Simulationstechnik, München 2024. 159-166.
BibTeX: Download