Apply an Edge AI Drop-In Solution to Enhance Wireless Condition-Based Monitoring

Condition-based monitoring (CbM) helps prevent equipment failures through predictive maintenance, but designing an effective system has typically required the optimal integration of precision sensing, low-noise signal chains, power management, and wireless connectivity. These are complex functions that can delay and increase the cost of CbM deployment. Designers are also recognizing the benefits of artificial intelligence (AI) analytics at the edge, which further complicates CbM. What is needed is a more straightforward and effective solution.

This article provides a brief overview of CbM. It then introduces a drop-in solution from Analog Devices that enables immediate deployment of wireless CbM with edge AI.

Why condition-based monitoring matters

Unplanned downtime remains a significant challenge in maintaining a high level of operating equipment effectiveness. A single unexpected failure in a critical piece of equipment can bring entire production lines to a halt, disrupt supply chains, and lead to expensive service interventions. Traditional maintenance approaches involving reactive repairs after a failure or rigidly scheduled service intervals have their downsides: reactive maintenance results in costly downtime, while scheduled maintenance incurs resource costs with needless replacement of components that remain operational.

CbM enables the implementation of more cost-effective predictive maintenance methods. By monitoring vibration, temperature, current, or other performance indicators, equipment operators can identify early warning signs of component degradation before failure occurs. This data-driven approach reduces unplanned downtime, extends equipment life, and lowers the total cost of ownership.

For all its benefits, CbM deployment can stall due to the complexity of the requirements and the need for expert knowledge across multiple disciplines. Overcoming these challenges presents a significant obstacle to the successful application of CbM-enabled predictive maintenance for industrial and automotive equipment.

Challenges and requirements in condition-based monitoring

For CbM to deliver its full potential benefits, CbM solutions must perform reliably in demanding industrial and automotive environments while delivering timely analysis based on accurate measurement data. Yet, the nature of these target environments places considerable mechanical and environmental stress on measurement devices even during normal operation of the monitored equipment. Industrial motors, drivetrains, and heavy rotating equipment expose monitoring devices to constant vibration, shock, temperature extremes, and high levels of electromagnetic interference (EMI).

To enable reliable predictive maintenance, vibration sensors in CbM devices must be able to detect the more subtle changes that often provide the earliest clues of shaft imbalance, misalignment, or bearing wear. Ensuring high precision vibration measurement despite harsh environmental conditions requires a high-bandwidth, low-noise sensor signal acquisition subsystem that delivers stable performance in challenging operating environments.

At the heart of CbM methods, vibration analysis provides the foundation for the recognition of patterns that differentiate normal operation from early indicators of failure. In the past, vibration sensor systems passed their measurements along to a central host or cloud-based resource for analysis. However, advanced CbM solutions have increasingly begun to shift analysis to the edge. By analyzing data within or near the sensor system, results are generated with minimum latency, and traffic is reduced on time-sensitive industrial and automotive networks.

In particular, edge AI inference based on convolutional neural network (CNN) models enables real-time interpretation of vibration changes. However, inference using CNN is computationally intensive, further complicating the goal of implementing CbM without exceeding system power, size, or cost limits.

The need to minimize power consumption becomes more acute as CbM finds increased use in rotating equipment or in remote or mobile equipment where wired connections are impractical. In meeting requirements for wireless connectivity in these cases, Bluetooth Low Energy (BLE) offers the required combination of range, power, and reliability compared to alternative connectivity options (Table 1).

Table 1: Among wireless connectivity standards, BLE offers a combination of characteristics suitable for wireless vibration monitoring. (Table source: Analog Devices)

As with edge AI processing, however, the challenge is finding a BLE connectivity solution able to operate within the power constraints of a wireless sensor system. In fact, ensuring extended battery life continues to be a challenge for designers of any wireless sensor system. However, it is particularly important in industrial and automotive applications where sensors might be difficult to reach. In a CbM system expected to perform CNN inference, both battery and power management become increasingly critical. The challenge here lies in orchestrating multiple regulators, sequencers, and charging systems to reduce power consumption while ensuring stable operation.

Evaluation kit provides a drop-in wireless CbM solution with edge AI

Analog Devices’ EV-CBM-VOYAGER4-1Z Voyager 4 kit addresses the challenges of deploying wireless CbM with edge AI by providing a complete battery-powered vibration monitoring platform for ongoing evaluation of CbM technology or immediate deployment in predictive maintenance applications. The kit is designed to withstand harsh environments, using a vertical standoff (Figure 1, top) that firmly holds the main printed-circuit board (pc board) in place on one side and a battery on the other. A power pc board and sensors are located at the bottom of the standoff, close to the vibration source to be monitored. For deployment, the vertical standoff assembly is placed in a protective aluminum enclosure (Figure 1, bottom) with a diameter of 46 millimeters (mm) and a height of 77 mm. The enclosure is topped with an ABS acrylic lid to allow BLE connectivity.

Figure 1: Voyager 4’s rugged standoff assembly and protective enclosure enable reliable wireless CbM with edge AI in harsh environments. (Image source: Analog Devices)

Built around an Analog Devices MAX32666 BLE microcontroller unit (MCU) and an Analog Devices MAX78000EXG+ AI MCU, the wireless sensor system design integrates a comprehensive set of low-power devices for delivering precision vibration measurement and anomaly detection with extended battery life (Figure 2).

Figure 2: By combining multiple low-power devices, Voyager 4 provides the combination of sensing, processing, and connectivity required for a drop-in wireless CbM edge AI solution. (Image source: Analog Devices)

For vibration measurement, the Voyager 4 uses Analog Devices’ ADXL382-1BCCZ-RL7 three-axis accelerometer, which combines microelectromechanical systems (MEMS) sensors, an analog front-end (AFE), and a 16-bit analog-to-digital converter (ADC). Featuring an 8 kilohertz (kHz) measurement bandwidth, this device is designed to deliver accurate measurements even in high-vibration environments. It is well-suited for low-power designs, consuming only 520 microamps (μA) in high-performance mode with 8 kHz bandwidth, or just 32 μA in a low-power mode with 400 Hz bandwidth.

In the Voyager 4’s system design, the ADXL382’s output passes to the Analog Devices ADG1634BCPZ-REEL7 CMOS switch, which the MAX32666 BLE MCU controls. The combination of this BLE MCU and an Analog Devices ultra-low power ADXL367BCCZ-RL7 MEMS accelerometer plays a central role in the Voyager 4’s operating modes (Figure 3).

Figure 3: The Voyager 4’s operating modes ensure efficient generation of training data and real-time inference, demonstrating how edge AI can support predictive maintenance without relying on cloud resources. (Image source: Analog Devices)

During training operations (path "a" in Figure 3), the MAX32666 MCU channels raw vibration data from the ADXL382-1BCCZ-RL7 for transmission to the user’s host system through the MAX32666 BLE radio or through the Voyager 4's USB connection. As discussed later in this article, this operating mode provides the training data needed to generate custom inference models underlying edge AI for CbM.

During anomaly-detection operations (path "b" in Figure 3), the Voyager 4's MAX78000EXG+ AI MCU uses its direct connection to the ADXL382-1BCCZ-RL7 to read raw vibration data and execute a custom inference model with its integrated CNN accelerator for anomaly prediction. If the inference results indicate the presence of an anomaly, the MAX78000EXG+ issues an alert, which the MAX32666 BLE MCU passes to the user for action.

If no anomaly is detected, the sensor enters sleep mode. In this quiescent state, the ADXL367BCCZ-RL7 accelerometer draws only 180 nanoamperes (nA) in motion-activated wakeup mode, triggering when vibration exceeds an adjustable threshold. When this motion-activated wakeup occurs, the ADXL367BCCZ-RL7 in turn wakes the MAX32666 BLE MCU, which initiates a new vibration measurement and inference cycle. This approach helps minimize power consumption during normal operation, restricting power-intensive BLE radio use to training sessions and anomaly alerts (Figure 4).

Figure 4: Motion-activated wakeup and selective use of the BLE radio help extend Voyager 4 battery life. (Image source: Analog Devices)

Effective power management is essential in a device intended to predict failures of critical machinery and equipment. Along with system-level power savings enabled through the Voyager 4's motion-activated wakeup operation, the Voyager 4 integrates an Analog Devices MAX20335BEWX+T power management integrated circuit (PMIC) to deliver the required voltage supplies. In addition, an Analog Devices MAX17262 fuel gauge monitors battery current and supports battery-life estimation. During the Voyager 4's various operating modes, the MAX32666 MCU can enable or disable individual MAX20335BEWX+T outputs to match specific power needs, further optimizing power consumption.

At the device level, low-power operation is a core feature of the individual devices used in the Voyager 4 kit. For example, the MAX32666 BLE MCU requires only 27.3 microamperes per megahertz (μA/MHz) when executing from cache at 3.3 volts; the MAX78000EXG+ AI MCU uses 22.2 μA/MHz (While loop execution) from cache at 3.0 volts with its Arm® Cortex®-M4 core processor active. Furthermore, both MCUs integrate a dynamic voltage scaling controller that further minimizes active core power consumption.

This combination of system-level and device-level power optimization effectively minimizes power consumption during the Voyager 4's various operating modes. In its normal anomaly detection mode, the Voyager 4’s power consumption is about 0.3 milliwatts (mW) with the sensor active once per hour, translating to as much as two years of battery life for a 1500 milliampere-hour (mAh) battery under typical conditions. In contrast, training mode requires extensive use of the BLE radio to transmit vibration data for use in model training and validation, resulting in power consumption over 0.65 mW (see Figure 4 again).

Training and deploying a vibration monitoring model for edge AI

Training CNN models has become a relatively straightforward process with the wide availability of suitable software tools. In training models for edge AI applications, however, the resource limitations of edge processors and MCUs have driven the development of more specialized tools created to optimize models for individual target devices. Analog Devices provides such tools in its AI on a Battery GitHub repository, which guides users through a documented workflow. Analog Devices breaks the model workflow down into a sequence of three stages and provides a dedicated GitHub repository for each (Figure 5).

Figure 5: A structured workflow with dedicated repositories of tools and instructions helps developers optimize CNN models for the MAX78000EXG+ AI MCU, enabling practical AI-driven CbM on power-constrained devices. (Image source: Analog Devices)

In the initial stage, the ai8x-training repository provides detailed, step-by-step instructions for preparing the work environment and performing training with the included train.py Python script. In the next stage, the ai8x-synthesis repository provides a similarly detailed set of instructions for setup and operation of the tools used to convert a trained model into C code.

A critical factor in achieving success in edge AI is understanding the capabilities and limitations of the target CNN execution environment. Contained within the ai8x-training and ai8x-synthesis repositories, Analog Devices includes a detailed tutorial to help developers understand the relationship between the CNN model implementation decisions and the capabilities of the MAX7800x AI MCU.

The final stage, documented in the software development kit repository, provides the instructions and tools used to develop firmware that embeds the inference model for the target MAX7800x MCU. After generating the firmware, users load it into the Voyager 4 via wired or wireless update. At this point, the user can connect with the Voyager 4 over BLE and issue commands using a Python graphical user interface (GUI) running on a Windows host. In normal run mode, the AI MCU performs inference as directed by the MAX32666 BLE MCU or automatically on wakeup.

Conclusion

Unplanned downtime due to equipment failure drives cost and risk. Although CbM can help reduce cost and mitigate risk through predictive maintenance, the design of suitable wireless sensor systems with analysis remains complex. Analog Devices’ Voyager 4 wireless vibration evaluation kit provides a drop-in solution that overcomes these challenges, enabling rapid deployment of predictive maintenance with precision sensing, efficient power utilization, wireless connectivity, and robust processing with edge AI.