IGADA-IoT framework improves sensor accuracy by 7% with automated data augmentation

A team of researchers led by Mingchun Sun has introduced IGADA-IoT, an automatic data augmentation framework designed to optimize energy consumption in wireless sensor networks (WSNs) for IoT devices. Traditional augmentation methods rely on a single data generator and empirically set quantities, failing to adapt to dynamic information gaps and ignoring sample heterogeneity. IGADA-IoT addresses these limitations with a hierarchical multi-generator collaboration and scheduling strategy (HMGCS) that intelligently allocates generated samples across multiple generators. It also includes an information gap-model performance joint evaluation and closed-loop method (IGMP-EC) to prevent under- or over-augmentation.

Experimental results demonstrate that IGADA-IoT improves the average accuracy of multiple downstream models by 7.27% compared to standard baselines, and outperforms advanced augmentation techniques by 8.67%. When compared to individual generator approaches, accuracy gains reach 7.24%. The framework was validated using public IoT sensor datasets from the UCR Archive and real-world deployments, confirming its accuracy and generalizability. By reducing the need for frequent high-frequency sampling while maintaining data quality, IGADA-IoT provides a practical path to extend sensor battery life in edge AI and IoT applications.