FedDQ: An Intelligent Federated Data Quality Assessment for Wireless Sensors Network

The widespread adoption of IoT devices and wireless sensor networks (WSNs) has revolutionized various domains leading to unprecedented data generation and connectivity. Thus, ensuring the quality of this data is crucial, as inaccuracies can pose major risks in mission-critical applications. Context awareness is also crucial in this regard since it enables tailored assessment approaches that account for spatial, and environmental factors as well as available resources.Many solutions have been proposed for the assessment of data quality in IoT and wireless sensor network environments, but these solutions often lacked adaptability and overlooked context awareness. Our previous work introduced BIGQA, a context-aware big data quality assessment framework that utilizes link prediction techniques to generate a data quality assessment plan based on a knowledge graph representing data contexts.In this paper, we propose FedDQ, a novel federated approach for context-aware data quality assessment in IoT and WSN environments based on the contribution proposed by BIGQA. FedDQ decentralizes the quality assessment process and leverages federated learning techniques to help ensure data privacy and distribute the workload more effectively across the network. Our solution was implemented using the Pytorch and Flower federated learning framework and was evaluated using the radiation sensors’ monitoring dataset provided by the Lebanese Atomic Energy Commission (LAEC-CNRS). It was able to generate a valid data quality assessment plan in a federated environment.