The “High-Accuracy, Real-Time Location Services” demo tracks emergency room patients in real-time, showing:
- Geolocation of patients as they pass through hospital zones
- Wait time of patients in each zone
- Real-time visualizations and alerting to ensure the most critical patients receive timely care
Today we go to see how Fujitsu and Striim can combine to provide real-time location monitoring for healthcare. While the demo is direct today, healthcare, a similar mechanism can be used to streamline manufacturing, modernized retail and automated monitoring in aviation and many other industries. The goal if this use case is to fully monitor patient visits. We want to be able to do fine zones and monitor wait time by zones. We need to be able to send alerts if a wait time is too high and take immediate action with an end result of reducing wait time. In this demo locators a place to rent a room. We have configured different zones of interest. I provide a tag to patients as they arrive. The location is continuously tracked and the time they spend in each zone recorded.
The locators continually send data by UDP, which is combined with zoned geometries. We have defined using point in polygon, spatial queries, pattern matching spots, users entering zones, and if they’re there for too long while windows and aggregate queries keep track of wait time. All of this information including real-time tracking is displayed on an interactive dashboard. The solution is created through a partnership with Fujitsu. The Fujitsu devices send the locations and the stream triggers workflow for the Fujitsu cloud base. From my process with all of the processing taking place on Fujitsu m two systems, the benefits to the customer include real time location monitoring of patients with alerts for waiting in one spot for too long. The wait times if continuously monitored and workflows can be triggered based on rules. We are simulating this use case in our lab. We have set up two locators in a room and we’ll be tracking these badges using the locators on the dashboard. You can see the location of each badge shown by the path, the zones that we have set up and a heat map that shows the most trafficked areas. As a budge anticipate zone, the zone changes color and the time in that zone and starts to be recorded. If a badge has been in the zone for too long, then we will send an alert which can trigger a workflow. We also send alerts if the overall wait time per zone or per badge is too long.
We’re showing the average overall wait time and also breaking it down by badge and zone. The backend processing the powers. This starts by reading data from locators. You can see this raw data here.
We pull out the necessary fields and we’d do a spatial match of this data against the zones. You can see the badge data here with an indicator of whether or not it is in a zone.
This information is used to track entry and exit of zones in order to change the zone color and also to calculate average wait times by badge and zone. We use pattern matching to calculate the total time spent in a zone and to send alerts if someone is in a zone for too long. The dashboard was also built using our product. All of the visualizations are power bi queries. Here’s the query for the tracking map. It has a simple drag and drop configuration. The wait time charts are also powered by a query with a similar configuration, which shows how easy it is to add new visualizations and build dashboards like these. So you can see how the combined power of stream processing and Fujitsu devices and hardware console real world problems. Please subscribe to our channel for more videos like this one.