Our Lab Director Featured in UppTalk: Multi-Robot Collaboration and Bio-Inspired Systems

Published on

We're excited to share that our lab director, Dr. Didem Gürdür Broo, was recently featured in Uppsala University's UppTalk series – a popular science seminar that makes cutting-edge research accessible to alumni and the general public. The 32-minute conversation, hosted by Anneli Björkman offered fascinating insights into our lab's work on multi-robot systems and the future of human-robot collaboration.

🎬Video: Watch the video on YouTube
📌Venue: Blåsenhus, Uppsala, SE, 2025

Our experience at the SciFest 2025 in Uppsala

Published on

Dive into our experience of how it was for us to participate in such a STEM festival and get to know what we prepared for the participants.

📌Venue: Fyrishov Hall D, Uppsala, SE, 2025

Crowdsourcing eHMI Designs: A Participatory Approach to Autonomous Vehicle-Pedestrian Communication (RO-MAN 2025)

Published on

We recently explored how crowdsourcing platforms can be used to design external Human–Machine Interfaces (eHMIs) for autonomous vehicles (AVs). These interfaces are key for helping AVs communicate clearly with pedestrians.

🪧Paper: [Proceedings are not published yet]
📌Venue: 34th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), Eindhoven, NL, 2025

Enabling Symbiosis in Multi-Robot Systems via MARL (ICPS 2025)

Published on

We present a CTDE value-decomposition MARL scheme that shares only battery state to induce symbiotic coordination in warehouse robots. In long-horizon delivery with charger contention, it cuts completion time by about 10.7% and reduces task and energy imbalance by about 13.8% compared with non-symbiotic baselines.

🪧Paper: Enabling Symbiosis in Multi-Robot Systems Through Multi-Agent Reinforcement Learning
📌Venue: IEEE 8th International Conference on Industrial Cyber-Physical Systems (ICPS), Emden, Germany, 2025

Investigating Symbiosis in Robotic Ecosystems (ICRAS 2025)

Published on

We introduce a symbiosis-guided reward shaping approach for MARL in heterogeneous multi-robot systems. On
high-dimensional tasks (ShadowHand object passing, mobile manipulation), mutualistic rewards improve training stability and reduce variance compared with standard rewards.

🪧Paper: Investigating Symbiosis in Robotic Ecosystems: A Case Study for Multi-Robot Reinforcement Learning Reward Shaping
📌Venue: 9th International Conference on Robotics and Automation Sciences (ICRAS), Osaka, Japan, 2025
💻Code: RewMARL

Optimal Gait Control for a Tendon-driven Soft Quadruped Robot by Model-based Reinforcement Learning (ICRA 2025)

Published on

We first learn a DNN surrogate of the robot’s dynamics from Simscape trajectories, then train SAC on the surrogate for speed, and finally run post-training on Simscape to close the sim-to-real gap. A parametric gait space (trot) and an inverse-kinematics mapping reduce actions from 12 motor commands to 4 gait parameters per step, cutting exploration of infeasible motions.

🪧Paper: [Proceedings are not published yet]
📌Venue: IEEE International Conference on Robotics and Automation (ICRA), Atlanta, USA, 2025
💻Code: KTH-MasterThesis