ML

A unified ML framework—Event-Automata + Physics-Embedded Neural ODEs + Neural Substitute Solver—for accurate, real-time inference of hybrid (continuous–discrete) dynamics with reliable Sim-to-Real transfer to edge hardware.

Sim-to-Real Edge Digital Twins that fuse event-aware physics with neural operators for sub-microsecond inference, online parameter self-calibration, and control integration on FPGA/MPSoC.

Network-aware multi-agent RL that scales decentralized secondary voltage control by truncating critics to κ-hop neighborhoods with provable approximation guarantees; validated up to 114 DGs.