Members can download this article in PDF format.
Connected devices aren’t necessarily intelligent devices. However, by integrating artificial intelligence (AI) with edge computing, data can be processed and analyzed directly on devices at the "edge" of a network—at the point where a device connects to the internet, such as sensors.
In this way, edge AI brings computational power closer to the data source. Decisions can be made on the periphery of the device instead of in the cloud, enabling faster real-time decision-making, lower latency, higher power efficiency, and improved robustness and data security.
Sponsored Resources
Engineers today are challenged to design systems that can make accurate, intelligent decisions in real-time. One example is the smart-home security camera. To detect whether an object in your backyard is a dog or a stranger using cloud AI may require sending the data up to the cloud, processing it, and then sending it back down so that a decision can be made. This would not only take too long, but it would require a great deal of power to accomplish.
When making decisions at the edge, highly integrated embedded devices run neural networks locally, requiring less power and allowing for improved security and privacy.
Higher-Precision Fault Detection
Lots of data can be utilized by edge AI. For instance, an OEM building a car will have a large amount of sensor-obtained data on how the car will behave. Edge AI allows you to take this data and create a function that will detect a potential fault or do preemptive maintenance. It can boost fault-detection precision by processing this data locally and in real-time, reducing latency and increasing responsiveness.
In applications such as motor drives and solar-energy systems, real-time fault detection ensures both operational safety and long-term reliability.
Edge AI-enabled MCUs can monitor two types of faults:
A motor-bearing fault occurs when there are abnormal conditions or deterioration in the bearings of an electric motor. Detecting these faults is vital to prevent unexpected failures, reduce downtime, and reduce maintenance costs.
An arc fault is a discharge that occurs when electricity flows through an unintended path. These are often caused by insulation breakdowns, loose connections, or other faults in the system. The discharge can generate intense heat, leading to fires or damage to the electrical system. Arc-fault detection will prevent fires and ensure safety.
Embedded Processors Increase Intelligence at the Edge
Texas Instruments offers a range of processors with varying processing capabilities, enabling developers to choose the right level of AI performance for their application while maintaining low power consumption. Integrated edge-AI based fault detection in a real-time MCU such as the TI TMS320F28P550SJ can help avoid false alarms while providing better predictive maintenance (Fig. 1).
