AWS brings machine learning solutions to take on manufacturing challenges

In December, 2020, AWS announced the general availability of Amazon Lookout for Equipment - a highly pragmatic solution that uses machine learning models to help customers perform predictive maintenance on equipment.

Amazon Lookout for Equipment is just one of a growing arsenal of robust solutions focused on using machine learning to solve real industry challenges.

We recently sat down with Jan Metzner, Head of Manufacturing for EMEA; and Constantin Gonzales, a Principal Solutions Architect working with customers in the manufacturing industry. Some of the strongest use cases in the sector focus on quality assurance and predictive maintenance solutions. This is an area AWS has seen growing traction for its machine learning products, using its characteristic ability to hone in on the cause of the challenge and develop a pragmatic solution, AWS endeavours to provide solutions to specific challenges and free professionals to work on higher-value projects.

Indeed, a major contributor to AWS’s early success and rapid growth was its ability to bring solutions that significantly improved the productivity and efficiency of customers - tackling problems and overcoming obstacles at the technical core of their work, so they could focus their efforts on their core competencies.

This process of solving challenges at the coalface has become a vital part of how AWS builds new services. According to Metzner, the way they develop solutions now is roughly broken down as 90% based on direct feedback from customers; and 10% on what AWS thinks customers will need (with this 10% based on AWS’s understanding of their customers: what they need but cannot quite articulate).

Perhaps one of the most compelling examples of this in action is Amazon Monitron - a service that solves a real challenge in and outside of the manufacturing space – specifically the difficulty to integrate sensors and the cost/complexity to build an ML-powered predictive maintenance solution based on those sensors.

For manufacturers, sensors transmit data that helps them understand their environment and support important initiatives, such as quality assurance and predictive maintenance. By collecting and analysing data such as heat and vibration patterns, professionals can detect abnormal behaviour in machinery which might indicate that maintenance is needed. But the challenge is that many legacy devices may not have sensors built in. And retrofitting them can be disruptive, expensive, and in some instances, impossible.

Amazon Monitron tackles this issue head-on. The service is a cost-effective out-of-the-box solution consisting of wireless sensors to capture vibration and temperature data, a gateway device to securely transfer data to AWS, the Monitron service to analyse sensor data with machine learning, and a companion mobile app to set up the devices and receive operating behaviour alerts. In practice, technicians can install everything and start monitoring equipment in minutes.

We can find another example in AWS Panorama, a service that provides computer vision services that help enterprise leaders solve a raft of challenges. These include use cases like automating quality control by detecting defects on a manufacturing line, or providing critical input to supply chain operations by tracking throughput or optimizing freight operations in transport yards. The solution can also be used to ensure industrial site safety by identifying when workers or vehicles are straying into dangerous, off-limit zones. Crucially for these use cases, because the solution can operate in low-connectivity environments, it enables the processing of the data at the edge allowing them to access real-time predictions in remote and isolated places.

According to AWS, as we see the rollout of 5G, many of these solutions will expand into new roles within the enterprise. The more reliable, secure, and low-latency network capability will also see the development of other solutions from AWS, as enterprises find other hurdles they need to overcome. The firm's AWS Wavelength solution is an example of this development taking place already; the infrastructure offering enables enterprises to tap into a growing network of Wavelength zones that offer organizations the opportunity to tap into AWS compute and storage services within the data centres of communications service providers collocated with the carrier’s 5G network; significantly driving down latency and enabling them to take full advantage of the technical benefits that 5G offers.

Manufacturing in Europe will become a key area of innovation for technology firms over the next few years, as enterprises in the space search for technological advantages and drive sweeping transformation projects to ensure they remain competitive. A significant part of this journey will see firm's search for new revenue streams and opportunities - such as offering greater customization enabled by smart factories; or digital services embedded directly into their after-products (such as intelligent maintenance alerts and performance insights). For AWS, this presents a growing opportunity to bring their solutions to market - and help enterprises cut to the core of challenges.

0 thoughts on "AWS brings machine learning solutions to take on manufacturing challenges"

Leave a Reply

Share via ...