On the Efficiency of LoRA Fine-Tuning for Vision-Language-Action Models in Industrial Robotic Manipulation - 2026-07-11 - Research Publication - The study compares LoRA and full fine-tuning for a flow-matching VLA on an exact physical UR5e setup. Four tasks cover cylindrical and hexagonal bolt insertion, distractor-aware bolt pick-and-place and bearing press-fit.… - source: arxiv.org
Domain Arithmetic: One-Shot VLA Adaptation under Environmental Shifts - 2026-07-01 - Research Publication - DART isolates reusable domain directions from one-shot fine-tuned VLA weight updates using subspace-aligned weight arithmetic. The UR5e is evaluated in MimicGen simulation for cross-embodiment transfer from a Panda robot.… - source: arxiv.org
A Multimodal Manufacturing Safety Chatbot: Knowledge Base Design, Benchmark Development, and Evaluation of Multiple RAG Approaches - 2025-11-14 - Research Publication - The authors introduce a multimodal chatbot powered by large language models that meets these design requirements. To evaluate their solution, the authors developed a domain-specific benchmark of expert-validated question… - source: arxiv.org
Using Knowledge Representation and Task Planning for Robot-agnostic Skills on the Example of Contact-Rich Wiping Tasks - 2023-08-27 - Research Publication - To achieve that in this case study, their approach needs to address different kinematics, gripper types, vendors, and fundamentally different control interfaces. The authors conducted the experiments with a mobile platform… - source: arxiv.org
Generalized Object Search - 2023-01-24 - Research Publication - The thesis formulates generalized three-dimensional object search as a partially observable decision process that exploits spatial structure and language. The robot-independent system is deployed on Boston Dynamics Spot… - source: arxiv.org
Learning to Play Cup-and-Ball with Noisy Camera Observations - 2020-07-19 - Research Publication - The study presents a learning model based control strategy for the cup-and-ball game, where a Universal Robots UR5e manipulator arm learns to catch a ball in one of the cups on a Kendama. High-fidelity Mujoco simulations… - source: arxiv.org