In a century of automotive progress that started with the Ford Model T, learning to drive a car has become an almost universal rite of passage, and good driving skills are a point of pride for many people (think Formula 1 or NASCAR, or that friend who parallel parks effortlessly). As a result, it often is difficult for us humans to imagine becoming just passengers in our own self-driving vehicles.
While the pandemic has emptied many roads and highways, authorities have seen a rise in the percentage of crashes among those still driving. We know that replacing or supplementing certain human tasks with data and computational power can improve outcomes. And, while the debate about autonomous vehicle safety continues, we can see the potential for safer roads from AI-enhanced driver assistance systems, for example.
The driving forces behind autonomous vehicle technology are computer vision (CV) and artificial intelligence (AI). Applications for these technologies exist in the utilities industry as well, and innovators are finding them. As corporate utility leaders evolve their business models to remain competitive by exploring new market opportunities through innovation or partnership, AI and CV offer promise for improving both supply chain and grid management.
Predicting the future
As an advanced subset of AI, “computer vision” refers to any method that allows machines to interpret visual input. Early approaches decomposed an image into a one-dimensional array of pixel data and used the borders and edges of shapes to render 3D models from 2D images.
CV has since evolved to feature-based object recognition and the use of convolutional neural networks, or CNNs. Social media sites offer a great example of the power of CNNs. When a user is tagged in an image, an algorithm “learns” the features of that user’s face. CNNs then apply a set of filters to compute the likelihood that a new image matches a certain user. Such methods also can help identify suspected shoplifters in real-time, reducing inventory loss.
As the use of imagery from satellites and drones becomes standard in utilities asset management, these same CV techniques can help predict when assets need to be repaired or replaced. In other words, CV and AI can help utilities predict the future.
Mining massive volumes of data—the “new gold”
Today, utilities collect massive volumes of data through sensor networks, smart meters, customer payment systems and satellite imagery. However, they need an effective way to mine this data from a myriad of underlying systems. They also need to organize the data into a structured network model that reflects all of their grid assets. By doing this, they can have a proper foundation on which to layer business applications that integrate CV, AI and other technologies like intelligent automation and data analytics to drive performance and boost efficiency.
Using CV and AI to improve asset management can decrease costs in the supply chain by reducing overstock and by predicting the true age of asset failure. This approach also can model and predict energy usage across variables such as time of day and weather, allowing utilities to supply the correct amount of power to the grid. The benefit is achieving the higher level of grid resiliency and reliability that utilities are striving for, and customers are growing to expect.
To overcome the underlying data challenges described above, our CGI OpenGrid360 solution suite provides an Integrated Network Model that enables the use of a variety of data sources—such as geospatial, imaging, and asset traits—to monitor, predict, and safeguard infrastructure health.
While there are clear applications for CV and AI in the utilities industry, striking the proper balance between speed and stability is essential.
Please contact me to explore the possibilities of CV and AI, enabled by CGI OpenGrid360 and our team of utilities industry experts.
Learn more at CGI.com.