Edge computing with Azure Percept.

We had an opportunity to be among the first in the world to try out Microsoft’s new Azure Percept development kit as part of the private preview, and we gave our best to explore possibilities with this new device.

Industry 4.0 is accelerating connectivity and automation, not the least within manufacturing. Connected things can signal their environment, be remotely monitored, controlled – and increasingly, make decisions and take different actions on their own. Basic use of the data generated from IoT devices could be used to trigger alerts. In case of excessive heat or vibration that the sensor detects, it can trigger an alert and a technician would be dispatched to check it out. In a more sophisticated IoT system, you might have dozens of sensors monitoring many aspects of industry operations. Those devices could learn how the system should work, detect different anomalies, adapt their behaviour to the changes and adjust their responses based on what they learn about the system.

By collecting the data from the connected machines, the production line can be visualised digitally and all data in the different parts of the production process could be visualised in real time. This offers the ability not only to monitor the production remotely but can also allow for simulations of adjustments before being implemented in the productions. Collection of digital data representing digital objects is called digital twin, and these virtual replicas of physical devices are often used in manufacturing and many other industries.

IoT sensors can produce vast amounts of data. Often bringing all the data to the cloud would be both too slow and too expensive to meet the requirements. Edge computing brings compute close to where the data is generated, allowing processing of data on the edge before important data points are sent to the cloud. By distributing workloads to the edge, devices not only spend less time communicating with the cloud, but they react more quickly to local changes, and even operate reliably in extended offline periods.

Azure Percept is a platform of hardware and services that simplifies use of Azure AI technologies on the edge. The development kit comes with an intelligent camera, Azure Percept Vision, and linear microphone array, Azure Percept Audio. Azure Percept works out of the box with Azure services such as Cognitive Services, Machine Learning, Live Video Analytics, and others to deliver vision and audio insights in real time. With Azure Percept you can quickly start and apply AI solutions to edge cameras and listening devices. Scenarios like object detection, spatial analytics, anomaly detection, keyword spotting, and others can easily be solved with use of pre-built Azure AI models for edge.

In order to test the Azure Percept development kit, we prepared an assembly scenario in which robotic hands place bolts at defined locations. With the use of AI and Azure Percept Vision we detect correctly placed and misplaced bolts and display real time information via a custom developed dashboard. In case of detected anomaly, a service case will be created in Dynamics 365 and notification is sent out. This notification could, for example, be sent to a technician to resolve the detected issue. With the use of Power BI dashboards, we show interactive reports of the data for specific time periods.

Information about Azure Percept development kit can be found here: http://aka.ms/getazurepercept. Azure Percept Studio, which allows you to connect, build, customise and manage your Azure Percept edge solutions is described here: https://aka.ms/azureperceptbuild. Azure Percept devices come with built-in security on every device and information about can be found here: https://aka.ms/azureperceptsecure. Azure Percept library of AI models is available here: https://aka.ms/azureperceptexplore.

There is enormous potential in edge computing and in Industry 4.0, it is already playing a lead role. These technologies can be applied in many ways in the industry, so if you would like to learn more about what business value you can get, don’t hesitate to reach out to us.