Java and IoT In Motion
At JavaOne last week, I was involved in the construction of the Internet of Things (IoT) concept project called "IoT In Motion," a joint partnership between Eurotech, Hitachi, Hitachi Consulting, and Oracle. I helped build some of the back-end components including a RESTful service written in Java with some database queries, and I helped a bit with the front-end as well.
Also known as "The People Counter", the IoT system included stereoscopic cameras that can detect shapes in three dimensions with movement. The software working with these cameras is programmed to know the difference between a human being and, say, a dog or cat, and report only when people have walked by, including the direction they were headed. Placing these cameras at the entrances and exits to various places within The Moscone Center, such as the keynote lecture halls, event rooms, and show floors, the traffic patterns of people could be analyzed as they walked in and out of these places.
And that's exactly what was done at Oracle OpenWorld and JavaOne this year. Software written in Java, end-to-end, made it possible to analyze traffic patterns over time, as well as registration line wait times. The GUI application that presented all of this data was built using JavaFX. The entire proof of concept was coordinated by Oleg Kostukovsky of Oracle, who you can see here explaining the system. Additionally, the folks at Eurotech and Hitachi posted these slides to explain the technology behind the system in detail. For a sample of the JavaFX GUI, check out this video.
With the system, Oracle was able to see, in real-time and historically, when people arrived for certain events, when they left and, and when correlated with the event schedule, draw conclusions as to why. For instance, once it was announced that Larry Ellison would not be able to make his presentation half-way through a keynote event, the IoT in Motion system reported just how much of an impact this news had on the room attendance: More than half of the people walked out within a few minutes time. This reaction was observed in real-time by the system coordinators, and drove home just how much value the analytics a system such as this, built almost entirely in Java end-to-end, could provide to event planners. Food for thought.