Structured Neural Network Dynamics for Model-based Control
This work addresses the challenge of efficient model-based control for autonomous systems, though it appears incremental by adapting existing linear dynamical system concepts to neural networks.
The authors tackled the problem of integrating learned system models with gradient-based model predictive control by proposing a structured neural network architecture inspired by linear time-varying dynamical systems, which simplifies analysis and control and removes the need for costly online derivative computations, demonstrating efficacy in standard continuous control domains.
We present a structured neural network architecture that is inspired by linear time-varying dynamical systems. The network is designed to mimic the properties of linear dynamical systems which makes analysis and control simple. The architecture facilitates the integration of learned system models with gradient-based model predictive control algorithms, and removes the requirement of computing potentially costly derivatives online. We demonstrate the efficacy of this modeling technique in computing autonomous control policies through evaluation in a variety of standard continuous control domains.