Wireless sensor networks (WSN) prove to be an enabling technology for Industry 4.0 for their ability to perform in autonomous manner even in regions of extreme conditions. Autonomy brings in independent decision making and exerting controls without manual intervention and frequent maintenance. This paper aims to inculcate intelligence to the WSN exploiting the merits of Artificial Intelligence (AI) algorithms in cheap and most preferred ESP8266 and ESP32 based nodes. Autonomy is brought in by means of optimal data transmission, compressive sensing fault detection and network reconfiguration and energy efficiency. Optimal data transmission is achieved using Q-learning based exploration exploitation algorithm. Compressive sensing performed using Autoencoders ensure reduction in transmission overhead. Fault detection is done using Binary SVM classifier and the network re-configures based on physical redundancy. This paper highlights the implementation of such autonomous WSN in real time along with their performance statistics.

中文翻译：

---

资源受限节点中基于AI的自重构和最优数据传输算法的自主传感器网络实现

无线传感器网络（WSN）被证明是工业4.0的使能技术，因为它们即使在极端条件下也能以自主方式运行。自治带来了独立的决策和施加控制，而无需人工干预和频繁维护。本文旨在在便宜且最受青睐的基于ESP8266和ESP32的节点中利用人工智能（AI）算法的优点向WSN灌输智能。通过最佳数据传输，压缩感测故障检测以及网络重新配置和能源效率来引入自治性。使用基于Q学习的探索开发算法可实现最佳数据传输。使用自动编码器执行的压缩感测可确保减少传输开销。故障检测使用二进制SVM分类器完成，并且网络基于物理冗余进行重新配置。本文重点介绍了此类自主WSN的实时实现及其性能统计信息。