An AI agent-based smart campus management platform

Abstract

The continuing digital transformation of higher education has produced large volumes of heterogeneous campus data, including network access records, wireless access-point mappings, academic information, security alerts, and institutional documents. These resources are often isolated across operational systems and therefore remain difficult to trace, integrate, interpret and use in a timely manner. This paper presents the design of an AI agent-based smart campus management platform intended to convert such fragmented data into a governed and explainable basis for campus operations and student services. The proposed work combines multisource data acquisition, layered data governance, Wi-Fi-based student trajectory reconstruction, spatiotemporal behavior analysis, anomalous traffic detection, interactive visualization, automated reporting, and conversational analysis. Its technical route separates transactional and analytical workloads through MySQL and ClickHouse, uses Kafka for stream ingestion and decoupling, employs Redis and scheduled jobs for responsive services, and introduces an AI Agent with retrieval-augmented generation (RAG) for tool-mediated data queries and knowledge-grounded responses. The platform is organized as a modular, independently deployable system for a single institution and incorporates role-based access control, desensitization, audit trails, and controlled knowledge publication. The expected value is a reusable architecture that supports campus situational awareness, interpretable student activity analysis, earlier risk identification, evidence-informed resource planning, and lower-barrier access to institutional analytics.