Machine condition monitoring and predictive maintenance are crucial technologies in modern industrial settings. Wireless sensor networks (WSNs) are commonly used to gather machine data with high flexibility and minimal installation effort. However, traditional WSN approaches that periodically or selectively transmit raw data either lack predictive capability or consume excessive energy. Furthermore, conventional static Edge-AI models running entirely on sensor nodes struggle to adapt to dynamic and complex industrial conditions due to limited labelled failure data and unpredictable machine dynamics. In this paper, we propose and evaluate a hybrid edge-central AI architecture. In this approach, sensor nodes perform the feature extraction as the first layer of the AI model, while deeper adaptive model layers operate at the central base station. This approach reduces energy consumption by limiting radio transmissions and enabling the use of complex, adaptive AI models. We validate the proposed architecture by implementing a set of common features on a typical ARM Cortex-M4 microcontroller used in wireless sensor nodes. We target the architecture of our previously developed wireless 1kS/s (kilo-sample per second) accelerometer. Results demonstrate that these features can be computed in only 32.5 ms and consume 32.43 μW. This represents a significant energy saving compared to raw measurement transmission (686.4 μW), highlighting the effectiveness and feasibility of our hybrid approach for industrial monitoring.

This work evaluates experimentally the applicability of wireless sensors networks (WSNs) based on Bluetooth Low Energy (BLE) for operational modal analysis (OMA) in machine diagnostics. OMA enables the dynamic properties of a mechanical system to be determined based on vibration measurements. A WSN for collecting such measurements requires precise synchronization, the transfer of relatively large volumes of data, and a lifetime of several days or weeks. The higher data rate of BLE 5 (2 Mbps), compared to other prominent radio technologies used in WSNs, offers several advantages, e.g. it could allow a reduction in the overall radio utilization, which generally contributes the most to the power consumption of a sensor node. However, the reliability of BLE 5 links in a machine environment must be assessed to determine the applicability of this radio technology in OMA. This work presents experimental results on the quality of BLE 5 links operating inside an industrial machine which indicate that such links can operate reliably when using a relatively high transmission power (above 0 dBm).

Selection of an energy efficient wireless transceiver is an important step in the design of any low power wireless system, specifically wireless sensor networks or IoT devices. In this paper, a figure of merit (FoM) for evaluating the energy efficiency of wireless radio transceivers is proposed. Using this FoM, it is possible to easily compare the performance of transceivers built with completely different technologies, in term of energy efficiency. This FoM can be used as a practical tool for researchers and engineers to evaluate wireless transceivers and determine the state-of-the-art or select a desired transceiver.

Wireless sensor networks (WSNs) are gaining increasing popularity in industry. Success of these systems depends on two very important factors: power efficiency of the sensor nodes and communication reliability. In this paper, we investigate the effect of using unidirectional (broadcasting) communication on the power consumption and reliability of a wireless sensor node. A high-precision wireless temperature sensor reported in an earlier publication is employed as a case study. First, we calculate the energy required to transmit and receive a message using Bluetooth low energy (BLE) in the physical layer without taking any reliability precautions. Then, we estimate the amount of energy required for the reliable transmission of a BLE packet using the BLE acknowledgment method and forward error correction (FEC) in the application layer. Through this paper, we show that the power consumption of a wireless temperature sensor can be reduced using broadcast communication and simultaneous forward error correction while providing enough reliability in short ranges.