TY  - EJOU
AU  - Hanzla, Muhammad 
AU  - Alabdullah, Bayan 
AU  - Shorfuzzaman, Mohammad 
AU  - Alonazi, Mohammed 
AU  - Almotiri, Jasem 
AU  - Jalal, Ahmad 

TI  - Edge-AI Enabled Multi-Agent Intelligence Framework for Evolutionary Optimization and Mobility Localization in 6G-IoT Smart Sensing Systems
T2  - Computers, Materials \& Continua

PY  - 
VL  - 
IS  - 
SN  - 1546-2226

AB  - The increasing adoption of wearable Internet of Things (IoT) devices and wireless sensor networks has accelerated the demand for intelligent, low-latency human activity recognition and localization systems. However, dynamic environments, sensor noise, motion variability, and communication delays continue to hinder reliable real-time sensing. To address these challenges, this study presents an Edge-AI enabled multi-agent intelligence framework for human activity recognition and mobility localization within a 6G-enabled IoT edge-cloud continuum. The proposed framework employs Self-supervised Transformer-based Adaptive Sensor Encoding and learning-aware temporal windowing to generate robust representations from heterogeneous sensing modalities. Extracted locomotion and localization features are optimized using the Aquila Optimizer (AO) and classified through a deep neuro-fuzzy inference model. Furthermore, an edge-assisted monitoring platform, EvolveSense, is developed to support real-time visualization and low-latency intelligent analytics. Experimental evaluation on the ExtraSensory and Opportunity datasets achieved activity recognition accuracies of 93.63% and 95.57%, respectively, while localization accuracy reached 91.33% and 92.40%. The results demonstrate the effectiveness of Edge AI, intelligent sensing, and optimized multi-modal learning for next-generation IoT healthcare and smart environment applications.
KW  - Multi-agent intelligence; Internet of Things (IoT); human activity recognition; mobility localization; edge computing; wearable sensors; smart healthcare; edge AI

DO  - 10.32604/cmc.2026.086599