From Sensors to Silicon: Causal Cross-Layer Self-Healing Edge AI for Secure and Energy-Adaptive IoT Systems

Impact 6 arXiv cs.AI

Summary

  • A new paper introduces a causal cross-layer self-healing Edge-AI framework designed for secure and energy-adaptive IoT systems.
  • Existing edge-intelligence approaches often treat dimensions like sensing quality, hardware reliability, and security in isolation.
  • The proposed framework continuously observes synchronized telemetry across sensor health, inference uncertainty, hardware counters, power states, and communication.
  • A temporally informed causal reasoning engine estimates the root cause of abnormal behaviors and quantifies diagnostic uncertainty during concurrent perturbations.

Why it matters

This research provides a unified approach to root-cause diagnosis in resource-constrained IoT nodes, helping distinguish between sensor failures, distribution shifts, hardware faults, and security threats.

Who should care