KVN1017: Multisentinel AGT Swarm Guard - Multi-Domain Autonomous Inspection Ecosystem

CHIN KAH MIN ASIA PACIFIC UNIVERSITY

i3DC26 | Tertiary (Online)

CR: 0.0000 | 0 Likes | 2 Views | 8 times | LS: 8.0
Like it? | Support them now!

Modern industrial facilities face critical inspection challenges including unplanned downtime costing up to USD 125,000 per hour, chemical safety incidents, and chronic worker exposure to hazardous environments. This paper presents SwarmGuard, a multi-domain autonomous inspection ecosystem comprising  specialized Autonomous Mobile Robots (AMRs) operating as a coordinated swarm or individually. HydroGuard, an explosion-proof chemical hazard detection platform featuring multi-gas sensor arrays for real-time chemical detection and localization and  PredictiveGuard, a predictive maintenance platform integrating thermal imaging and acoustic signature analysis for early equipment degradation detection. The core innovation lies in the integration of swarm intelligence algorithms with vision-language models (VLMs) through a central AI orchestrator that enables distributed autonomous operation, collaborative SLAM-based mapping, and multi-perspective sensor data fusion. Natural language processing capabilities automate report generation, contextual decision-making, and human-robot interaction. The system employs dynamic task allocation with fault-tolerant role redistribution, ensuring continuous mission-critical operation. Targeting semiconductor fabrication, EV assembly, and pharmaceutical manufacturing sectors, SwarmGuard demonstrates a projected ROI of 400–800% over five years with significant reductions in safety incidents and operational downtime, positioning it as a scalable Industry 4.0 solution for next-generation autonomous facility management.