What is the Internet of Things?
Image by Simon Zhu via Unsplash
Our new Martian friend just landed on Earth and is excited to learn about the latest developments in human technology. In this Q&A series, IBM experts explain complicated topics to a Martian (and you).
IoT, IoT, IoT! It’s all over your Earth news.
Thomas Bauer is IBM Watson IoT Industry Lab Leader for Government. I asked him to explain the Internet of Things to me.
Greetings! What is this Internet of Things I keep hearing about?
Greetings from Munich. Have you visited us yet?
Just peeked down from outer space.
Beam down so that we can show you around.
The Internet of Things is about connecting sensors embedded in everyday objects like cars, washing machines, helicopters, and street lights with the internet. These objects are enabled to send and receive data so that they can ‘talk’ with each other, with the user, the developer, or the maintenance company technician. Because more and more objects today have embedded sensors that produce loads of data, we need to find smart ways of analyzing this huge amount of data.
Why would you want to analyze that data?
Analyzing the data allows us to discover new ways of solving problems, saving costs, and making better and faster decisions. Ultimately, analyzing the data helps us find new ways of making life easier for people and companies.
Technology lets us understand and extend the insights gained from IoT data. Without that technology in place to analyze all the data in real-time, its value is minimal. The information’s usefulness would be limited by its complexity and scale.
I’ve also heard a lot about artificial intelligence. Does that mean you’re simulating human brains?
Well not exactly, but we do teach our systems to think. IBM Watson Cognitive Services can create and explore its own hypotheses, understand natural language, or detect emotions.
Take the sentence, “The Golf is parked in the garage.” Watson can understand that we’re talking about a car, not about the game of golf. It does that by surfing the rich digital rivers of IoT data, and other structured and unstructured data inputs. And that means that Watson can uncover patterns and insights that were unattainable before.
Systems like Watson actively learn from data, context, and interactions — and they evolve over time to continually improve and refine their responses and recommendations. Watson can give recommendations to decision makers and, where appropriate and allowed, implement actions such as safety controls, energy management, quality assurance, and other essential tasks.
This Watson sounds important.
Yes, and getting better and better. Some say Watson is growing up — like a child who’s continuously learning.
What do you mean by unstructured versus structured data?
Structured data shows up in a fixed format. You find it in spreadsheets, for example, where the output data of sensors like temperature, humidity, or speed can be collected.
Pictures, videos, tweets, text messages, and spoken language are all examples of unstructured data.
Unstructured data comes, from example, from a camera installed on a drone, taking photos of roofs after a thunderstorm. These photos are very interesting for insurance companies who want to know what claims costs they can expect. The Watson Visual Recognition solution can help classify the type and degree of damages from hail on a roof. With machine learning, the system gets smarter with each new photo it evaluates.
How do you teach a system?
You teach the system the difference between a damaged and undamaged roof by showing it photos of the undamaged roof and then comparing photos of that same roof with hail damage from different perspectives. Based on this input, the system is then prepared to classify roofs with hail damage. This can help insurance companies get a quick overview of the damage volume they can expect.
Visual recognition has many further possible applications, in areas such as manufacturing, retail, and education.
What else can you do with all that data?
You can correlate, combine, and re-combine the data, find patterns and come up with new insights and recommendations. For example, you can analyze historical data of car accidents in a city, correlate them with real-time traffic and weather data, and come up with risk assessments on the street segment level. This can help city command centers prepare first responder resources like police, emergency rooms, and fire stations.
There are so many applications and possibilities!
Yes! Some of the cases we’re working on in the Watson IoT Center in Munich might have a large impact on our society. Our researchers are currently testing a new technology of touchless — or cuffless — blood pressure acquisition, utilizing camera systems in the lower band of the infrared spectrum, which monitors your body at the capillary level. That can open a broad variety of new use cases, which are not limited to the medical area. It could even change disaster management and first responder’s actions.
What’s the biggest potential for IoT? What inspires you about it?
What inspires me is that our journey is just starting. Where does IoT — and especially IoT in combination with supercomputing and cognitive analytics — take us? We’re just beginning to ask the right questions, and we don’t have all the answers yet.
But we have so many opportunities to find completely new answers to these questions in our IoT ecosystem together with our clients, partners, start-ups, and research institutes.
Wow, that got deep. And fascinating! Thank you.
I think so too. You’re welcome.