Networked objects are learning to anticipate our needs and orchestrate responses that deliver safety, efficiency and convenience.
Ever had one of those days where everything goes wrong right from the start? Maybe you accidentally turned your alarm off instead of hitting snooze, and now you have to rush out the door for work. Of course the city would be doing road work today and, in this traffic, 15 minutes late is suddenly looking more like 40.
Compounding problems like these may be avoided as the world becomes more connected.
Sensor-embedded technologies interpret real-time data about their surroundings, which enables them to activate at just the right moment (or even before) with increasingly relevant responses. Sensors are in just about everything now — put them all together and you’ve got a coordinated network of objects delivering safety and efficiency. As they continue to collect data over time, these sensors will evolve from reactionary to anticipatory. Working together, they have the potential to save us a lot of grief.
To learn more about this trend, we talked with Jennifer Healey, research scientist at Intel, and Tim Plowman, embedded user experience lead with Intel Labs’ Experience Design team.
Where can we expect to see changes as a result of this trend, and what would that look like?
Jen: We’ve been looking a lot at anticipatory technology in the transportation space. In particular we developed a Concierge System, which was designed to anticipate when you needed to get up to get to work. For example, if traffic or weather was going to bad, or if there was some sort of indication that your commute might take longer than expected, the system was could wake you up a bit earlier so that you would have the additional time you needed.
This anticipatory alarm received information from the cloud regarding weather and traffic, but also from the car itself. The car could also push a message to the cloud that its gas level was low. The system would then anticipate that the driver might have to stop for gas and add that to the expected commute time.
We’re also looking at real-time anticipatory interaction between cars and smart infrastructure. A smart traffic light could let you know it’s about to turn red, and your car could tell the traffic light, “Look, I’m approaching.” With that, the traffic light could look both ways and, if if the coast is clear, change green for you so you could go through.
Tim: We started out primarily focusing on the in‑vehicle context. We were asking, what are the anticipatory experiences that could be delivered within the vehicle itself?
The vast majority of trips that we take in vehicles tend to be trips we’ve made before, but having that information, creating a profile or history in order to derive conclusions about what someone might be doing on a Tuesday at 5:30, provides useful information. If it’s aggregated and people opt in, that can beneficially impact the traffic load balancing. It could help create an efficient use of infrastructure and help the overall impact of transportation as it plays out.
With more connected ecosystems, where cars talk with one another, how do you see this playing out, or will we see this more in cities which are more predictive in nature?
Tim: It’s complicated. There’s a need for alignment at a regional, state, municipal and federal level for this to happen. There’s a big [privacy and security] issue around people being willing to share their routing pattern, being willing to submit that data to analysis and aggregation. Vehicles symbolically have represented autonomy and independence.
If you are going to do the things we’re talking about, you have to participate in collective activity.
You can step back and say, “Well, look at what everyone does with social media, you’re giving up huge amounts of privacy.” As an independent individual, an anticipatory automation experience in a vehicle has some benefits. When you’re leaving to drop your kid off at school, your history or your profile knows there are three routes that you typically drive to get your kid to school. It also knows that Tuesday is trash day, or it knows that this route is clogged for whatever reason.
Even just that individualized anticipatory automation experience has value. The real advantages come when many, many hundreds of thousands of cars are involved as opposed to one individual experience. The data sharing piece is a big hurdle.
Jen: Data sharing can either be voluntary, like sharing your location data with mapping apps, but it’s just as likely to be involuntary, such as when your car is tracked by law enforcement, tax enforcement or traffic control authorities.
This sharing is a negotiated agreement. People agree to share their data in exchange for the benefit of knowing the traffic conditions on a map app and allow the government to monitor them in exchange for the benefit of knowing that laws are being enforced to make the roads safe for everyone.
People are willing to give up some aspects of privacy and control in exchange for a perceived benefit. When it comes to autonomous driving, I believe that people will be willing to exchange their position, velocity and route data in exchange for the benefit of not having to drive, especially in high-traffic situations when driving is not really all that fun.
We can look beyond fleets of cars and think about entire city-wide ecosystems. How can weather affect different kinds of behaviors in a city?
On an overcast day, vendors are not going to sell as much ice cream and people are more likely to congregate in one part of the city versus another part of the city. Entire cities can start anticipating activity and demand based on weather. Looking at what the pedestrian shopping patterns are, and letting business know to expect either more or less business depending on their location. Cities can adapt parking, resources, and they can adapt things like where the police or first responders need to be to anticipate that, on a hot day, maybe more people are going to need medical attention.
Based on changing conditions and past things, we can grow a collective intelligence for the city. We can get the right resources ready to cope with things which we know generally happen under those conditions. That’s what you are looking for around resources.
How will these sensors fit into our daily lives?
Jen: We focused on older people, monitoring them and anticipating their daily patterns. If they generally got up and made tea, and then one day they didn’t, that could be cause for concern just based on normal habit.
Another thing that we looked at was medication reminders. We would know when a person needed to take their medication, then we would track their motion through the house. We could send a reminder to their phone that it was time to take their medication.
We’ve also looked at are smart lighting and smart heating. If somebody starts walking in the house from their bedroom to the bathroom, we can automatically turn on those lights and maybe start heating up the boiler.
Tim: The challenge is certainly regulatory. HIPAA regulations around the collecting data related to health and showing that data are basic hurdles. But the essential value in applying consumer data to that area really goes back to understanding behavior, and being able to extrapolate interventions based on that behavior.
If you measured the distance of one room to another, you could analyze a person’s gait based on the way they walked down the hall. You could use that as a baseline to understand whether or not an individual, maybe in their 80s, is having a good or bad day. Was their gait normal or was their gait slow or unstable? Were there pauses down the hall?
The challenge is finding the right to individual markers to track, and best behavior markers to interpret.
Cultural aspects of aging are very different in the United States compared with Asia and Europe. It’s challenging to ensure that what you’re doing is actually helpful, and not having some unintended consequences.
Jen: There’s this idea of anticipatory decisions. Should a system take action without checking with you or should it ask you first?
The right answer actually has to do with how much you trust the system.
When your car detects that you are in a crash and there is no time, you want the air bag to automatically deploy without checking with you first. However with something like grocery ordering, if your fridge notices you’re out of a certain thing, should it immediately reorder it or should it check with you first?
I think we’re going to go through some growing pains where we’re going to have to have a human in the loop for a while. At what point does the anticipatory automation become good enough so that people don’t mind actions being taken for them without their final check?
In the Real World Web, iQ by Intel and PSFK Labs explore the role Internet-enabled technologies will play in connected ecosystems of the future. This series, based on a recent report, looks at the rise of the Internet of Things and its impact on consumer lifestyles.
The Promise Of Predictive Computing: Anticipating Your Every Move
Are We Ready for Information Openload?
Why ‘Access’ Will Define the Internet of Things
Why ‘Contexual’ Is the New ‘Search’ Living Within the Internet of Things
The Feel-Good Effects of Affective Computing
The Real-Time Sensor Networks Building a Complete Picture of Our World
Uncorking Distributed Intelligence Networks at a Global Scale
Subscribe to iQ by Intel and stay up on innovation everywhere via Flipboard.