Early IoT Applications Illustrate Emerging Trends

Chung-Sheng Li
September 9, 2014


The emerging Internet of Things (IoT) is often discussed as a phenomenon of the future, rather than as an enabler of current applications. Yet early applications in use today can provide a clear sense of the shape of things to come, if not the full gamut of possibilities.

Too often, in my view, the “full gamut of possibilities” blinds us to the IoT’s present capabilities. The notion that, if we connect a zillion “things”, applications will emerge may be true. But a more pragmatic approach would be to ask, which things should we connect, and to what end?

In this article, I’ll describe a few characteristics of the emerging IoT and then turn to several specific, current applications to bring the concepts, use cases and challenges into sharper focus. We shouldn’t lose sight of the fact that the purpose of connecting “things” is to aid our quest to understand the world around us in order to better adapt to it. Achieving this end will require an incremental, practical approach: domain-specific applications will arise first, and data science will connect them

Vertical domains, horizontal ties

The IoT, in my view, represents an interdisciplinary quest to understand and anticipate real-world behavior to provide us with a more efficient, convenient, safer, even anticipatory world.

How will this actually work? Here’s a vastly simplified picture: sensors will provide the data to construct and populate behavioral and/or structural models that reflect the world around us. Missing data will be interpolated from existing data based on the models. Using advanced processing capabilities, these models will then extrapolate the data to anticipate future behaviors and outcomes and simulate “what if” scenarios to assess various options for action. Actuators will carry out the desired actions. Eventually, with continuous iterations, these processes and models will become more sophisticated, enabling more accurate reflections of reality and longer-range predictions of our world’s likely behaviors. This, in turn, will create more viable and diverse options for our adaptive responses.

In domain-specific cases, such as weather forecasting, public health management and urban challenges such as traffic control, safety and crime prevention, which we’ll discuss here, these concepts are already being applied to achieve valuable outcomes.

IoT behavioral models for various vertical domains, such as the aforementioned examples, will eventually be connected, improved and expanded by the use of “horizontals,” i.e., domain commonalities such as the emerging field of data science. What can we learn when multiple domains are combined or juxtaposed and their causal relationships identified? Early IoT applications provide examples of possible outcomes.

I can offer one IoT mantra in this regard: the better the data, the better the model; the better the analysis, the better the predictions. The volume of data is less important than constructively leveraging the best data to build a better model.

The business case

Before we examine several IoT use cases, we should recognize that, in the short run, economics likely will drive the search for positive business cases in vertical domains. In fact, that effort is already underway as various companies across a diverse set of industries seek to monetize commercial advantages in this emerging field.

Yet using the IoT for the common good is a laudable goal. In some cases, IoT applications may be exempt from the pressure to determine a return on investment (RoI), particularly where the common good outweighs the need for a positive business case. As we shall see, mitigating disease outbreaks certainly falls in that category.

Real-world IoT applications

Now let’s consider three real-world IoT applications, from the familiar to the emergent, to put these ideas on a practical footing.

Weather prediction is a good example of a familiar IoT application that’s already widely in use. Sensors on the surface, in the ocean, in the air, in the outer atmosphere and in space all provide data on the variables that affect the weather. At first these weather models merely described present conditions. As they improved, however, we’ve been able to predict the weather many days in advance with a relatively high degree of confidence.

Another application with real-world consequences: predicting disease outbreaks. In 1993, the “Four Corners” area where Colorado, Utah, Arizona and New Mexico meet experienced an outbreak of Hantavirus Pulmonary Syndrome (HPS). Since that outbreak in 1993, about 606 people have been infected and about 36 percent of those people died over the past 20 years. HPS is a horrible disease; a patient in late stage HPS often exhibits shortness of breath as their lungs fill with fluid.

Scientists from several institutions including Johns Hopkins University and IBM worked together on a behavior model that can assess the potential risk of such an outbreak occurring at a certain location. The starting point of the model is climate-influenced weather. In this case, El Niño typically introduces heavy rain into the arid Southwest of the USA, causing substantial vegetation growth. Subsequently, La Niña reduces or stops the rainfall and desiccates the vegetation. Vegetation provides cover for rodents and, when the vegetation disappears, rodents have nowhere to hide but in human habitations. As it turned out, rodents are the primary carrier of hantavirus. The behavioral model developed for the spread of hantavirus depended on combining the weather model, the vegetation model and the rodent population model to gauge the impact on a structural model of where people actually live.

The result was a forecast that predicted areas with a high risk of a future outbreak of hantavirus with high confidence many months in advance.

Because many ailments fall into the vector-borne disease category – including malaria, West Nile encephalitis and Dengue Fever – this same methodology has been applied to build behavioral models for other, similar health threats as well. As our use of IoT concepts becomes increasingly sophisticated, we may develop other applications that offer advanced risk assessment for other disease outbreaks.

Big city traffic and crime control

Other IoT applications are more prosaic. To reduce inner city traffic, London has been using surveillance cameras to track and impose charges on vehicles entering the downtown area within the London Inner Ring Road since 2003. Crossing an invisible line into downtown triggers an automatic snapshot and recognition of a vehicle’s license plate. The system bills the car owner instantaneously and payment is required by end-of-day or a severe penalty is imposed. Stockholm also has implemented a similar approach since 2007.

In yet another application, New York City has surveillance cameras in downtown and midtown to detect, for instance, people who leave a bag unattended. In Los Angeles the police use surveillance cameras on their cars to scan the license plates of parked cars. If the license plate raises a red flag the police can take pre-emptive action.

So the IoT already has been used for weather, disease outbreaks, traffic control and crime fighting, even anti-terrorism – just a few applications among many in use or being developed.

Additional issues

This initial essay on the IoT obviously leaves out many topics deserving of more attention. Data science and closed-loop system design will continue to be subjects of concerted enquiry. Models and simulations must become more sophisticated. We need to explore the specific technology improvements in sensors, centralized and distributed processing and actuators that will be needed to advance the IoT. On the public policy front, what are the IoT’s implications for data security and privacy? The list of topics for further discussion is long.

At least one simple conclusion is warranted here. The futuristic-sounding IoT depends on many well-known technologies already in use. Thus the IoT is at once more tangible and practical than its name suggests, yet its applications remain largely unexplored and under-exploited.



Chung-Sheng LiChung-Sheng Li is currently the director of the Commercial Systems Department. He has been with IBM T.J. Watson Research Center since May 1990.

His research interests include cloud computing, security and compliance, digital library and multimedia databases, knowledge discovery and data mining, and data center networking. He has authored or coauthored more than 130 journal and conference papers and received the best paper award from IEEE Transactions on Multimedia in 2003. He is both a member of IBM Academy of Technology Leadership Team and a Fellow of the IEEE.




2014-10-30 @ 8:47 AM by Holland, Steven

Will you be posting your slides from your excellent webcast yesterday on ",

Orchestrating the Smarter Planet in the world of the Internet of Things"?

2014-10-30 @ 3:09 PM by Ayer, Renee

Hi Steven, Yes, we will share it on the IEEE IoT web portal:  http://iot.ieee.org/ along with a recording of the webinar.