Home  >  Article  >  Technology peripherals  >  Bringing AI to the edge: Focus on the pipeline first

Bringing AI to the edge: Focus on the pipeline first

PHPz
PHPzforward
2023-04-09 17:31:011281browse

Data pipelines and pipelines that support edge AI and machine learning capabilities are critical.

Bringing AI to the edge: Focus on the pipeline first


Are artificial intelligence and machine learning starting to play a role in edge areas? Gradually, yes. But enterprises still need to focus on the underlying data pipelines and pipelines to support artificial intelligence and machine learning capabilities. Rohit Kadam, product manager at Mitsubishi Electric Power, said: "Whether you use different microservices and how you want to deploy or use them, you should pay more attention to building its architecture. Once you have the data, focus on how to connect the pipelines."

Kadam participated in a recent panel discussion on the opportunities and challenges of edge implementation, describing how the company's battery power plants and systems are connected through IoT and edge systems, allowing companies to monitor their health and expenses. Kadam explained that at Mitsubishi Electric, artificial intelligence and machine learning play an important role in notifying the company of issues within the connected battery power packs it delivers to customers, as well as managing downstream IoT devices. "The way ML works is we learn the behavior of batteries, so we know how much charge is in those batteries, or how much range is left. Those are some of the key metrics we use when training models. The more we learn, the better."

Through its combined strengths and artificial intelligence capabilities, Kadam said: “We now have the ability to look ahead in terms of data to identify and make decisions about safely trying to operate these plants. If it is confirmed that red flags are seen, there A built-in safety mechanism that starts up and then shuts down the plant in an orderly manner if necessary. This is already built into our solution and from an edge computing perspective, the distributed architecture can help take action in real time."

Operational metrics ensure battery system availability and assurance. “We have IoT usage metrics to track the properties of batteries and how they degrade over time. We think of the batteries themselves as edge nodes or edge computing devices. It helps track and monitor the battery’s voltage, current and temperature. . We process it and store the information there and then transmit it north back to the history server."

There are a lot of parts in the supply chain that need to come together, which makes standards an issue," Kadam said. There are no unified standards there, just trying your best to comply with various standards related to battery power plants,”

The challenge is that “battery power plants themselves are a unique space.” Kadam continued. This is a market serving electric vehicles, power grids, substations and building automation systems. "We have blend canvases and try to blend them together and then stream them up north. We actually parse all of that data and blend it together to more efficiently flow the different data sets back to the historical server."

The above is the detailed content of Bringing AI to the edge: Focus on the pipeline first. For more information, please follow other related articles on the PHP Chinese website!

Statement:
This article is reproduced at:51cto.com. If there is any infringement, please contact admin@php.cn delete