The MIT Enterprise Forum has an RFID and Sensors Circle that has monthly sessions about RFID, Mobile or sensor-related topics, often the latest amazing new device or locating technology or advancements in sensors. While these are remarkable developments, the real questions are, “Now that we have the ability to collect all of this fantastic data, what do we actually do with it? Where’s the value?” To help answer these questions, ChainLink and MIT hosted an event called “Smart devices? How About Smart — Really Smart — Clouds?”. Turns out, a lot of people want to know these answers — and turned out for this sold-out event.
Continuous Decision Intelligence
Three guest speakers on a panel were moderated by Ann Grackin. First up, Dennis Groseclose, President of TransVoyant, a company that has built a real-time decision platform for the industrial internet (VoyantCDI) that enables spatially oriented, time-critical decision-making in a scalable cloud environment. This “Continuous Decision Intelligence” (CDI) capability is derived by ingesting and processing massive flows of disparate data streams, and applying complex, multi-dimensional business logic. These data/event flows can include sensor data from machines, internet-of-things data, twitter feeds, and just about anything else you can throw at it. The platform executes rules to tell you when a relevant confluence of events has occurred.
The U.S. government uses this platform extensively for situational awareness applications, high end military intelligence, evaluating structured and unstructured data (e.g. video feeds) to quickly decide in real time what actions should be taken given huge numbers of inputs: here’s a known ‘bad actor’ near another bad actor; the weather is right to take action, but changing fast; they are not near a school or crowded area; so now is the time to take action.
Democratization of Sensor Data
Dennis said, “All the world is a sensor.” Now we have RFID, live video, unattended ground sensors, satellite, and so much more. TransVoyant felt it was important to democratize these capabilities. They designed their system so that rules can be written by the user with the domain knowledge and need, rather than having to be translated and written by some programmer. He gave as an example, oil company workers in Nigeria who were able to describe under what conditions the risk of people bunkering (stealing) the oil was the highest: according to the day of the week, the temperature outside, specific sensor readings, etc.. Because VoyantCDI gave them (the people with the domain expertise) the ability to write, try out, and adjust these rules, the users were able to describe and capture the behavior of the thieves and help thwart more theft.
And it is designed to be configurable by and useful to average citizens as well. Dennis gave the example of setting up the system to automatically send a message to his son, when several conditions are met: e.g. when air quality is X, pollen counts over Y, and/or other conditions are met, it sends an alert to tell him to bring his inhaler to school.
The Time-Value of Information
This becomes really valuable when there are important, high consequence, high value decisions that have to be made quickly — where you have compressed decision cycles, requiring analysis of complex, unstructured, multieam data. Dennis made the point that this sensor data and other data streams are living and changing and flowing like a river (not static like data sitting in a data warehouse). It gets stale quickly; includes both structured and unstructured data; is usually messy; is usually spatial, and temporal, and contextual.The temporal dimension really matters. Like if you are looking for drug dealers, you might look for when a plane is flying at a low altitude, and a boat crosses the path travelled by the plane, within a 6 hour window, adjusting for the direction and speed of the ocean currents, all in real-time.
The range of applications for this technology is very broad — from Industrial Internet players to hedge fund managers. For example, LiveSafe, whose founder was shot at the Virginia Tech shooting, has created a crowdsourcing security system to give rapid alerts and crime awareness on college campuses.They use TransVoyant to add alerts for weather, natural disasters, floods, and other events.
Critical Item Deliveries
Next up was Mike Lee, CEO of Airclic, provider of a cloud-based platform for managing last mile, critical item deliveries. Delivering perfect orders — the right items, undamaged, delivered on time, packed correctly, with accurate, complete documentation, and invoiced correctly — is the ideal for just about any company that delivers products. Depending on the criticality of the items though, some companies need to be ‘more perfect’ than others. This is where Airclic helps. They are tracking about 10M items per day, or $165B in business transactions per year. Airclic’s platform captures data from multiple points throughout the last mile supply chain, with a focus on performance.
They capture orders from upstream in the Supply Chain and provide control over the distribution network and the items in the network. Through scanning or reading RFID, they can ensure the right items are loaded in the right sequence in the truck. They can re-optimize deliveries based on the changing circumstances on the ground, what is actually happening throughout the day, including things traffic. They can manage by exception and, with live data, take proactive steps. They can provide the early notification when things are running late, so corrective action can be taken so that service providers and customers are not blindsided.
Transformation — From Paper-Based to Automation
For any company that has logistics challenges, it is transformational to move from primarily paper based management systems which are error prone, where things go missing, leading to lots of disputes and dysfunction. One of Airclic’s customers is US Foods, the second largest food distributor in the US. In California, they are required by law to take the temperature of certain foods (like frozen chicken) before they are delivered.Airclic’s workflow detects whenever the driver is in California and delivering frozen chicken, then a form pops up requiring them to take a temperature reading before they are allowed to deliver. Another long-term Airclic customer, Staples, happens to be the second largest internet retailer after Amazon. Airclic automates Staples’ huge operations to make sure those internet orders are handled properly and on time.
Another Airclic customer, a 3PL, delivers auto parts to General Motors’ dealerships. These are time critical, high-value deliveries. Airclic’s platform receives ASNs from GM, so the 3PL knows what’s expected and when. When line haul trucks deliver items into the cross dock, they are scanned off the truck, compared against the ASN, scanned again as loaded onto the outbound truck, and completing the loop with a scan when delivered to the car dealer. This complete and accurate chain-of-custody, with a reliable electronic Proof-of-Delivery (ePOD) saved Airclic’s customer about $100K per month from car dealers claiming they never received items.
Healing Healthcare
Finally we heard from Trevor Gruby, CTO and co-founder of Intelligent InSites a real-time operational intelligence platform for healthcare including asset management, patient workflow, environmental monitoring, infection control, and more. Trevor pointed out that hospitals have 20%-30% more equipment and supplies than they actually need. For example, when a nurse can’t find an IV pump, it is very stressful. So nurses stow away a couple extra so they know they will always have one. This leads to overbuying and underutilization, as well as difficulty in tracking that the proper device cleaning and maintenance tasks are done at the required time.
In many hospital settings, there are just too many manual processes, like recording the temperatures for vaccine stage units, answering phone calls in the OR to update family members on the status, updating the bed management system, searching for an infusion pump, entering ER/OR milestone time stamps, or calling housekeeping to request room cleanings.One nurse said ‘I spend half my career looking for things and the other half filling out paperwork.’ Clinicians are supposed to be there to help people, not spending all this time on these non-value add activities. It is no wonder we have the most expensive healthcare in the world!
Intelligence Starts With Interactions
Trevor said “intelligence starts with interactions.” Like automatically monitoring the temperature of the vaccine at the vaccine fridges. Intelligent InSites deploys literally tens of thousands of sensors in a hospital to achieve this. By doing this, they are able to dramatically reduce errors, and save a lot of time for everyone. They are able to make things run much more smoothly and efficiently and reduce wait times. In addition, this creates a treasure trove of very fine-grained data about actual operations, which can be mined to look for ways to continually improve performance.
The platforms are smart enough to interpret the incoming data properly, from many sources and formats. They took care to build flexibility at each step, including a flexible rules engine to interpret all that data; whether it is structured or unstructured.
Smart Platforms Extract Value from Rivers of Data
From all three of these speakers, we saw how these ‘rivers of big data’ being generated by the ever-expanding world of sensors, RFID, and locating devices, can be converted into high value by these really smart cloud platforms. This way RFID and sensor technology can really realize their full potential.
To view other articles from this issue of the brief, click here.