arXiv:2606.05045v1 Announce Type: cross Abstract: We present in this paper a framework for the identification of control-affine reduced-order models (ROMs). The proposed method utilizes autoencoders (AEs) to transform the high-dimensional states, and potentially the high-dimensional inputs, into reduced latent ones suitable for control-affine state-space dynamics. This is achieved by simultaneous training of the AE and the state-space model. In addition, we extend the discrete ROM formulation to a sequence-based model, which processes state and input histories to improve prediction accuracy wh
The proliferation of complex high-dimensional systems across engineering and scientific domains necessitates more efficient model reduction techniques to enable real-time control and simulation.
This development offers a novel approach to create accurate, control-affine reduced-order models, which is crucial for advancing AI's application in systems requiring high computational efficiency and real-time decision-making.
The ability to simultaneously train autoencoders with state-space dynamics for control-affine systems provides a more robust and integrated method for model reduction, potentially leading to widespread adoption in control systems.
- · AI/ML researchers and developers
- · Control systems engineering sector
- · Autonomous systems developers
- · Simulation and modeling software providers
- · Traditional high-dimensional simulation techniques
- · Systems requiring extensive manual model simplification
Improved performance and efficiency in complex control systems through more accurate reduced-order models.
Accelerated development and deployment of autonomous systems in critical infrastructure and industrial applications due to enhanced real-time control capabilities.
Potential for new classes of AI-driven control applications in areas currently constrained by computational complexity, such as advanced robotics and smart grids.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG