AI-based Framework for Robust Model-Based Connector Mating in Robotic Wire Harness Installation accepted at CASE 2025

Our paper “AI-based Framework for Robust Model-Based Connector Mating in Robotic Wire Harness Installation” (, ) has been accepted for presentation at the IEEE International Conference on Automation Science and Engineering (CASE) 2025.

In this work, we introduce a novel framework that integrates visuotactile learning with force control and first-order optimization to enable robust, fast, and certifiable connector mating in robotic wire harness installation. The system combines multimodal trajectory transformers with a differentiable shadow program, permitting environment-aware search strategy optimization directly on industrial robot controllers. We demonstrate substantial improvements in cycle time and success rates across diverse connector types, validating our method in automotive center console assembly.

Paper website: https://claudius-kienle.github.io/AppMuTT

Blog post:

Authors:
Claudius Kienle¹ ², Benjamin Alt³ ², Finn Schneider⁴ ², Tobias Pertlwieser⁴, Rainer Jäkel², Rania Rayyes⁴

IAS Lab, Computer Science Department, TU Darmstadt, Germany ↩
ArtiMinds Robotics, Karlsruhe, Germany ↩ ↩² ↩³ ↩⁴
AICOR Institute for Artificial Intelligence, University of Bremen, Germany ↩
Institute for Material Handling and Logistics (IFL), Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany ↩ ↩² ↩³