Architecture and Implementation of a Mobile Internet of Things-Enabled Logistics Management System

Abstract

The deep integration of the Mobile Internet of Things (MIoT) and edge intelligence enables intelligent logistics to enter a new phase of development. However, traditional end-cloud architectures are constrained in highly dynamic logistics scenarios, impeding large-scale industrial upgrading. Therefore, a three-level innovation framework encompassing paradigm, methodology, and enabling technologies was established, with proactive collaborative edge intelligence agents as the core paradigm to overcome the limitations. A hierarchical architecture consisting of an intelligent terminal layer, a collaborative edge layer, and a cognitive cloud brain layer was constructed, with three key methodological components integrated: dynamic federated task offloading, large language model (LLM)-driven decision generation, and privacy-enhanced cross-domain knowledge transfer. Correspondingly, enabling technologies—including a mobility-aware scheduling algorithm and an LLM-agent tool invocation mechanism—were designed. The study advances the deep integration of edge intelligence and MIoT technologies in vertical industries and exhibits substantial academic significance and industrial application potential.