文献类型：期刊文献

英文题名：Adaptive Agent-Based Image Scrambling for Scientific IoT Edge Sensing Using Hybrid Chaotic Maps, Fuzzy Control, and Swarm Intelligence

作者：Li, Yongjiang[1]

机构：[1] Guangdong Ocean University, School of Mathematics and Computer Science, Guangdong, Zhanjiang, 524088, China

年份：2026

外文期刊名：IEEE Internet of Things Journal

收录：EI(收录号：20263921532205)、Scopus(收录号：2-s2.0-105051188910)

语种：英文

外文关键词：Adaptive control systems - Authentication - Chaotic systems - Ciphertext - Cryptography - Internet of things

外文摘要：Scientific Internet of Things (IoT) edge sensing requires adaptive image protection under dynamic latency and energy constraints on resource-limited devices. This paper proposes an adaptive fuzzy–swarm hybrid-chaotic scrambling (AFS-HCS) agent that observes image, edge-resource, network/protection-demand, and feedback states. A hierarchical Sugeno policy and resource-bounded cooperative swarm refinement are introduced to adapt finite-precision hybrid chaotic parameters and hybrid-granularity scrambling. The proposed framework supports lightweight scrambling and standard authenticated transmission. Experiments on five public datasets compare AFS-HCS with image-scrambling methods, learning-based edge-decision frameworks, and standard cryptographic baselines, and quantify scrambling effectiveness, ciphertext-level statistical characteristics, and edge efficiency. The authenticated-transmission mode achieves ciphertext entropy of 7.9987, NPCR of 99.628%, and UACI of 33.463%, while the scrambling-only and authenticated-transmission modes require 17.2 and 21.2 ms, respectively, on Raspberry Pi 5. These results characterize the protection–latency–energy tradeoff under the evaluated conditions but do not establish cryptographic security for the scrambling layer. ? 2014 IEEE.