IPAB Workshop-25/03/2021 Gabriella Pizzuto Title: Physics-guided neural networks for dynamics model learning Abstract: Robotic manipulators interact with objects present in their surroundings. In such scenarios, model-based controllers would greatly benefit from having accurate and robust dynamics models. Whilst there has been numerous research efforts in learning such models, it is often the case that these frameworks are mainly tested in continuous motion, yielding poor results when deployed on real systems exposed to non-smooth frictional discontinuities. Inspired by a recent promising direction in machine learning, our work focuses on a newly developed methodology of learning dynamics models undergoing external impact. In this talk, I will present our current and ongoing work in this domain, with our contribution on using physics-guided neural networks for augmenting data-driven deep models with physical consistency. Lei Yan Title: Decentralized Ability-Aware Adaptive Control for Multi-Robot Collaborative Manipulation Abstract: Multi-robot collaboration is extremely challenging due to the different kinematic and dynamics capabilities of the robots, the limited communication between them, and the uncertainty of the system parameters. To address these challenges, we propose a Decentralized Ability-Aware Adaptive Control method, in which the force capability of each robot is maximized by exploiting its null-space motion, while the designed adaptive controller enables decentralized coordination according to the capability of each robot. Simulation results show the proposed method can achieve online adaptation and accurate trajectory tracking irrespective of the low-level controllers, and can be used for heterogeneous multi-robot systems. Mar 25 2021 13.00 - 14.00 IPAB Workshop-25/03/2021 Gabriella Pizzuto, Lei Yan Blackboard Collaborate
IPAB Workshop-25/03/2021 Gabriella Pizzuto Title: Physics-guided neural networks for dynamics model learning Abstract: Robotic manipulators interact with objects present in their surroundings. In such scenarios, model-based controllers would greatly benefit from having accurate and robust dynamics models. Whilst there has been numerous research efforts in learning such models, it is often the case that these frameworks are mainly tested in continuous motion, yielding poor results when deployed on real systems exposed to non-smooth frictional discontinuities. Inspired by a recent promising direction in machine learning, our work focuses on a newly developed methodology of learning dynamics models undergoing external impact. In this talk, I will present our current and ongoing work in this domain, with our contribution on using physics-guided neural networks for augmenting data-driven deep models with physical consistency. Lei Yan Title: Decentralized Ability-Aware Adaptive Control for Multi-Robot Collaborative Manipulation Abstract: Multi-robot collaboration is extremely challenging due to the different kinematic and dynamics capabilities of the robots, the limited communication between them, and the uncertainty of the system parameters. To address these challenges, we propose a Decentralized Ability-Aware Adaptive Control method, in which the force capability of each robot is maximized by exploiting its null-space motion, while the designed adaptive controller enables decentralized coordination according to the capability of each robot. Simulation results show the proposed method can achieve online adaptation and accurate trajectory tracking irrespective of the low-level controllers, and can be used for heterogeneous multi-robot systems. Mar 25 2021 13.00 - 14.00 IPAB Workshop-25/03/2021 Gabriella Pizzuto, Lei Yan Blackboard Collaborate