[{"abstract":[{"lang":"eng","text":"Deep reinforcement learning (RL) policies for robotic control are typically trained for a single hardware configuration. When a joint locks, the learned mapping no longer matches the available actuation. This paper investigates whether combining self-attention with explicit kinematic observations can improve a policy’s adaptability to varying active-joint configurations. We train a single RL agent to control a Franka Emika Panda manipulator across different sets of active joints, from a restricted 4-degree-of-freedom (DOF) mode to the fully redundant 7-DOF configuration. Instead of fixed-length state vectors, the method processes the active joints as a variable-length sequence, enriching each joint’s representation with its relative spatial offset and geometric Jacobian influence. The proposed kinematics-informed attention architecture reaches a success rate of 82.1% across trained configurations and 66.6% on unseen configurations, outperforming multilayer perceptron and generic attention baselines while remaining nearly collision-free. Deployed on a physical Panda, the simulation-trained agent reaches 79.0% success on the unseen configurations, closely matching the simulation on the same target schedule (77.0%)."}],"language":[{"iso":"eng"}],"user_id":"244831","date_created":"2026-09-21T16:34:09Z","publication_status":"accepted","status":"public","date_updated":"2026-10-21T09:44:11Z","publication":"9th Iberian Robotics Conference (ROBOT) ","author":[{"id":"244831","orcid_put_code_url":"https://api.orcid.org/v2.0/0009-0009-6800-1273/work/227375532","full_name":"Mayer, Patrick Thomas","first_name":"Patrick Thomas","orcid":"0009-0009-6800-1273","last_name":"Mayer"},{"id":"245736","full_name":"Rexilius, Jan","orcid_put_code_url":"https://api.orcid.org/v2.0/0000-0002-4579-214X/work/227375533","last_name":"Rexilius","orcid":"0000-0002-4579-214X","first_name":"Jan"}],"department":[{"_id":"102"}],"citation":{"ieee":"P. T. Mayer and J. Rexilius, “Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms,” in <i>9th Iberian Robotics Conference (ROBOT) </i>, Barcelona.","mla":"Mayer, Patrick Thomas, and Jan Rexilius. “Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms.” <i>9th Iberian Robotics Conference (ROBOT) </i>.","alphadin":"<span style=\"font-variant:small-caps;\">Mayer, Patrick Thomas</span> ; <span style=\"font-variant:small-caps;\">Rexilius, Jan</span>: Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms. In: <i>9th Iberian Robotics Conference (ROBOT) </i>","short":"P.T. Mayer, J. Rexilius, in: 9th Iberian Robotics Conference (ROBOT) , n.d.","ama":"Mayer PT, Rexilius J. Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms. In: <i>9th Iberian Robotics Conference (ROBOT) </i>.","bibtex":"@inproceedings{Mayer_Rexilius, title={Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms}, booktitle={9th Iberian Robotics Conference (ROBOT) }, author={Mayer, Patrick Thomas and Rexilius, Jan} }","apa":"Mayer, P. T., &#38; Rexilius, J. (n.d.). Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms. In <i>9th Iberian Robotics Conference (ROBOT) </i>. Barcelona.","chicago":"Mayer, Patrick Thomas, and Jan Rexilius. “Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms.” In <i>9th Iberian Robotics Conference (ROBOT) </i>, n.d."},"_id":"7175","type":"conference","year":"2026","keyword":["Reinforcement Learning","Robotic Manipulation","Kinematic Adaptation","Sim-to-Real Transfer"],"project":[{"name":"Institute for Building Intelligence","_id":"A827C0AA-C7DA-11E9-B0AE-1F4CB252D58D"}],"quality_controlled":"1","title":"Adapting to Variable Active-Joint Configurations via Kinematics-Informed Attention for Robot Arms","conference":{"end_date":"2026-11-20","name":"9th Iberian Robotics Conference (ROBOT) ","location":"Barcelona","start_date":"2026-11-18"}}]