In modern grids, renewable inverters lack the physical inertia of traditional turbines. My research develops Agentic AI to act as Synthetic Inertia, providing sub-cycle grid stability to prevent cascading failures and system collapse.
The core insight: treat the grid not as a passive network but as a dynamic multi-agent arena — where security, communication, and physical safety are inseparable. Aegis-Grid is the framework that operationalises this insight.
|
|
|
| Version | Research Question | Status |
|---|---|---|
 |
Fail-Secure Communication — minimal deterministic semantics, fragmentation & strict acceptance | ✅ COMPLETE |
 |
Thresholding Under Pressure — constrained adaptation under sustained adversarial conditions | ✅ COMPLETE |
 |
Autonomous Security Agents — local decision-making, fog-of-war, cost-aware mitigation | ✅ COMPLETE |
 |
Grid-Aware Security — safety policy derived from physical power system constraints | 🔵 ACTIVE |
 |
Post-Cryptographic Model — architectural stress-testing beyond classical crypto assumptions | 🟣 CONCEPTUAL |
| 🔐 | Fail-Secure First | Security evaluated by how it fails, not how it performs under ideal conditions |
|---|---|---|
| 📐 | Explicitness Over Convenience | All assumptions declared — hidden defaults are design errors |
| ⚙️ | Physical Correctness > Network Correctness | A message correct at network layer but harmful to the grid is invalid |
| 🔍 | Reproducibility as Primary Goal | Results are for inspection, questioning and reproduction — not marketing |
© 2026 Dr. Burhan U Din Abdullah · MIT License · IISc Bengaluru