Sensor calibration has emerged as a critical challenge in large-scale Wireless Sensor Networks (WSNs), where low-cost an‎d heterogeneous devices are increasingly deployed for environmental monitoring, smart agriculture, industrial automation, an‎d urban infrastructures. Although these sensors enable affordable an‎d scalable sensing, they suffer from drift, measurement bias, an‎d degradation over time, which compromise data quality an‎d system reliability. This seminar investigates the integration of data mining an‎d machine learning methods into network design to enable adaptive, secure, an‎d scalable calibration mechanisms. A multi-layered research framework was developed, combining simulation environments, federated learning, Gaussian Process Regression (GPR), AutoML pipelines, an‎d blockchain inspired security modules to ensure trustworthy an‎d energy-efficient calibration across dynamic contexts. Experimental results demonstrate that data-driven calibration strategies reduce mean absolute error (MAE) by over 38% compared to traditional techniques, while achieving robustness against drift, noise, an‎d adversarial conditions. Case studies across urban air quality monitoring, agricultural field sensing, an‎d industrial emissions compliance validate the framework’s generalizability an‎d practical applicability. The proposed system further incorporates adaptive scheduling, edge–cloud orchestration, explainable AI modules, an‎d real time dashboards, enhancing interpretability, scalability, an‎d user trust. This research contributes to the development of autonomous, resilient, an‎d sustainable sensor infrastructures, aligning with smart city, climate resilience, an‎d Industry 5.0 paradigms. Future directions include integrating neuromorphic edge intelligence, cross-domain transfer learning, an‎d open-source calibration frameworks to democratize access to intelligent sensor calibration an‎d advance global data integrity stan‎dards.

10/28/2025 12:00:00 AM

87985

1404/08/07

Study on data mining methods for sensor calibration in network design

(كاليبراسيون حسگر، داده‌كاوي، يادگيري ماشين، شبكه‌هاي حسگر بي‌سيم (WSN)، رگرسيون فرآيند گاوسي، يادگيري فدرال، AutoML)

(Sensor calibration, data mining, machine learning, Wireless Sensor Networks (WSNs), Gaussian Process Regression, federated learning, AutoML, edge computing, smart cities.) , (Sensor calibration, data mining, machine learning, Wireless Sensor Networks (WSNs), Gaussian Process Regression, federated learning, AutoML,)