Among the many challenges of parallel kinematic manipulators, achieving high-speed and accurate control remains crucial. Estimating their dynamic properties is essential for designing precise and efficient control schemes. Conventional methods for dynamic model identification have been effective, though deep learning approaches have historically faced limitations due to data inefficiencies. However, recent advancements in physics-informed neural networks (PINNs) offer a way to improve both control and the extraction of interpretable physical properties from these robots. In this work, we propose and validate a PINN-based dynamic model for a Delta parallel robot, specifically the ABB IRB 360-6/1600. Our approach incorporates known physical properties, such as mass matrix sparsity, to improve accuracy and computational efficiency in dynamic model identification. To the best of our knowledge, this is the first study applying PINNs to model parallel robots. The method is validated experimentally, and its performance is compared to a validated identification technique for physically consistent identification, demonstrating the effectiveness of this approach for real-world applications in parallel robots.

The determination of the dynamic properties of a robot is especially important for designing highly accurate and efficient control systems. Conventional methods for dynamic model identification have proven to be effective, where deep learning (DL) approaches have shown limits due to data inefficiencies. However, thanks to novel physics-informed DL architectures, such as Deep Lagrangian Networks (DeLaN) [1], it is possible to control and extract interpretable physical information of a robot. This paper introduces an augmented DeLaN architecture for linear viscous and non-symmetrical Coulomb friction identification, which also learns motor parameters such as rotor inertia. An approach is proposed for comparing this method with the conventional dynamic identification and previous DeLaN implementations. Moreover, our friction and rotor inertia identification is validated, and the performance of our model is analyzed with a real robot (UR5e).