Abstract
Urban resilience requires rapid deployment of monitoring systems capable of detecting radiological threats in transport networks without disrupting normal operations. This paper presents a multi-agent AI framework designed to accelerate the development and integration of isotope detection systems within smart city ecosystems. The proposed framework employs a hierarchical orchestrator-executor architecture in which specialized AI agents collaborate to generate synthetic training data via GEANT4 Monte Carlo simulations of NaI(TI) scintillation spectra, automate the development workflow including script generation, test-driven development, and result aggregation, and prepare a data foundation for deploying classification models at urban transport checkpoints. The modular architecture with a TCP API interface enables seamless integration with existing smart city platforms and scales for metropolitan deployment. This work demonstrates that heterogeneous LLM-based multi-agent orchestration substantially accelerates radiological simulation pipelines and establishes the direction for autonomous radiological monitoring at smart city transport nodes.

This publication has 24 references indexed in Scilit: