Discrete-time neural networks guarantee stable robotic dynamics modeling
Learning Dissipative Dynamics with Dissipativity-by-Construction Discrete-Time Neural Networks
Robotics
Summary
Models that predict how robots move need to be stable and safe to work well with control systems. The authors created a new way to train computer models that learn robot movements directly in discrete steps, ensuring these models always behave in a stable and safe manner. Their method avoids complex math during training, runs faster, and keeps the safety property built into the model instead of adding it later. Tests on robot simulations show their approach predicts movements as well as other methods but with better speed and guarantee of stability.
What this means in practice
- •For robotics engineers: Build robot control models that inherently guarantee stable and safe behavior in discrete time without computationally costly continuous-time solvers.
- •For control systems developers: Integrate learned discrete-time dynamics models with guaranteed incremental dissipativity to enhance feedback loop reliability in automation systems.
Authors
Tuan Luong, Hyungpil Moon
Abstract
Dissipativity is a fundamental system-theoretic property closely related to stability, passivity, and input--output stability, and is particularly important in robotics, where learned dynamics models are often embedded within feedback control loops. However, most existing approaches for learning dissipative dynamics are based on continuous-time formulations, which require ODE solvers during training or inference and can therefore be computationally expensive. Moreover, because practical implementations are inherently discrete-time, direct discretization of a continuous-time passive system does not necessarily preserve passivity, motivating the need for explicit discrete-time guarantees. This study proposes a method for learning incrementally dissipative dynamics from input--output time-series data using a deep multilayer perceptron formulated directly in discrete time. Through a constrained parameterization and a dedicated training procedure, the proposed model guarantees incremental dissipativity by construction rather than through regularization. Lyapunov-based analysis establishes the corresponding dissipativity and stability guarantees, while simulations on robotic dynamical systems demonstrate competitive prediction accuracy, computational efficiency, and consistent preservation of incremental dissipativity compared with baseline methods.