MuViS: Multimodal Virtual Sensing Benchmark

arXiv (Cornell University)

Abstract

Virtual sensing aims to infer hard-to-measure quantities from accessible measurements and is central to perception and control in physical systems. Despite rapid progress from first-principle and hybrid models to modern data-driven methods research remains siloed, leaving no established default approach that transfers across processes, modalities, and sensing configurations. We introduce MuViS, a domain-agnostic benchmarking suite for multimodal virtual sensing that consolidates diverse datasets into a unified interface for standardized preprocessing and evaluation. Using this framework, we benchmark established approaches spanning gradient-boosted decision trees and deep neural network (NN) architectures, and show that none of these provides a universal advantage, underscoring the need for generalizable virtual sensing architectures. MuViS is released as an open-source, extensible platform for reproducible comparison and future integration of new datasets and model classes.

Authors 8

1. Jens U. Brandt
2. Noah C. Puetz Aachen

   TH Köln - University of Applied Sciences · Leiden University

   Affiliation as printed

   Leiden University , Netherlands

   TH Köln , Germany

3. Jobel Jose George Aachen

   TH Köln - University of Applied Sciences · Leiden University

   Affiliation as printed

   Leiden University , Netherlands

   TH Köln , Germany

4. Niharika Vinay Kumar
5. Elena Raponi Aachen

   Leiden University

   Affiliation as printed

   Leiden University , Netherlands

6. Marc Hilbert Aachen

   Leiden University · Toyota Motor Corporation (Switzerland) · Laboratory of Racing Chemistry

   Affiliation as printed

   Leiden University , Netherlands

   Toyota Racing , Germany

7. Thomas Bäck Aachen

   Leiden University

   Affiliation as printed

   Leiden University , Netherlands

8. Thomas Bartz-Beielstein

   TH Köln - University of Applied Sciences

   Affiliation as printed

   TH Köln , Germany