DATA-DRIVEN nonlinear dynamical systems

Complex control theories are utilized when predicting human motion. To estimate human motion without using conventional control theories, deep learning can be used for human motion prediction. Therefore, an own-designed wearable device is created to collect data on human motion. By utilizing this collected data, motion prediction algorithms can be made to find the optimal number of joints and positions for effective prediction. [김동연]

Understanding differential equations directly from the physics law of nonlinear dynamical systems is a challenging problem. Recent studies, such as Sparse Identification of Nonlinear Dynamical Systems (SINDy), suggest a novel method that can automatically generate symbolic equations for a nonlinear coupled dynamical system directly from time-series data. [조정래]