Yang Xing, Shu Lei, Li Kailiang, Nurellari Edmond, Huo Zhiqiang, Zhang Yu
College of Engineering, Nanjing Agricultural University, Nanjing 210031, China.
College of Artificial Intelligence, Nanjing Agricultural University, Nanjing 210031, China.
Sensors (Basel). 2023 Jul 25;23(15):6672. doi: 10.3390/s23156672.
The Solar Insecticidal Lamp Internet of Things (SIL-IoTs) is an emerging paradigm that extends Internet of Things (IoT) technology to agricultural-enabled electronic devices. Ensuring the dependability and safety of SIL-IoTs is crucial for pest monitoring, prediction, and prevention. However, SIL-IoTs can experience system performance degradation due to failures, which can be attributed to complex environmental changes and device deterioration in agricultural settings. This study proposes a sensor-level lightweight fault-detection scheme that takes into account realistic constraints such as computational resources and energy. By analyzing fault characteristics, we designed a distributed fault-detection method based on operation condition differences, interval number residuals, and feature residuals. Several experiments were conducted to validate the effectiveness of the proposed method. The results demonstrated that our method achieves an average F1-score of 95.59%. Furthermore, the proposed method only consumes an additional 0.27% of the total power, and utilizes 0.9% RAM and 3.1% Flash on the Arduino of the SIL-IoTs node. These findings indicated that the proposed method is lightweight and energy-efficient.
太阳能杀虫灯物联网（SIL-IoTs）是一种新兴模式，它将物联网（IoT）技术扩展到农业电子设备。确保SIL-IoTs的可靠性和安全性对于害虫监测、预测和预防至关重要。然而，由于故障，SIL-IoTs可能会出现系统性能下降的情况，这可能归因于农业环境中复杂的环境变化和设备老化。本研究提出了一种传感器级轻量级故障检测方案，该方案考虑了计算资源和能源等实际约束条件。通过分析故障特征，我们设计了一种基于运行条件差异、区间数残差和特征残差的分布式故障检测方法。进行了多项实验以验证所提方法的有效性。结果表明，我们的方法平均F1分数达到95.59%。此外，所提方法仅额外消耗总功率的0.27%，并在SIL-IoTs节点的Arduino上占用0.9%的随机存取存储器（RAM）和3.1%的闪存（Flash）。这些发现表明所提方法是轻量级且节能的。
