Type-2 fuzzy logic empowered trajectory prediction for wireless sensor networks

Wireless sensor networks (WSNs) are increasingly utilized for object trajectory identification and tracking, especially in scenarios where global positioning system (GPS) or radio frequency identification (RFID) technologies are unavailable. Trajectory extraction in such networks presents significant challenges due to the need to analyze spatio-temporal sensor data and address uncertainty in trajectory points. To overcome these challenges, we propose a multilevel object-tracking algorithm powered by Interval Type-2 (IT2) Fuzzy Logic System (FLS), an Artificial Intelligence (AI) approach designed to represent and manage uncertainty effectively. Our methodology leverages the capabilities of IT2-FLS to accurately generate and predict object trajectories, even in the presence of noisy or incomplete data, by emulating human reasoning and decision-making processes. Comparative analyses confirm that our IT2-FLS-based approach outperforms the T1-FLS alternative by requiring fewer rules, achieving superior results in trajectory reconstruction and prediction tasks. Furthermore, the evaluations demonstrate the robustness of the proposed system in anomaly detection and predictive analytics, outperforming leading classification algorithms. The method not only delivers high accuracy but also preserves model interpretability and reduces rule complexity, which are critical for efficient operation in WSNs. These findings establish the IT2-FLS-based methodology as an innovative and effective AI-driven solution for trajectory prediction in WSNs, combining enhanced accuracy, robustness, and practical applicability in complex and uncertain environments.