The Rise of the GEOINT of Things
By Stuart Blundell, Gabe Chang, David Foster, and Michael Hauck
This article was originally published in USGIF’s State & Future of GEOINT Report 2017. Download the full report here.
The Internet of Things is one of the 10 elements comprising the GEOINT Revolution as described in the 2016 State of GEOINT Report. Arguably, it has been with us for about a decade now, but is still in its infancy. Even less mature, but just as inevitable, is the GEOINT of Things (GoT) being born out of the Internet of Things (IoT).
What is the GEOINT of Things? Imagine you have a trillion agents running around the planet doing your bidding. The Internet of Things is the current world of billions (and soon, trillions) of interconnected things that sense, think, act, and communicate. These things exist physically somewhere, so they have a location in “geospace.” IoT is inherently geospatial, so this is a huge opportunity for GEOINT practitioners to expand the domain of their tradecraft, especially when we consider what some of these “things” are. They can be as discreet as a tire pressure monitor on a car, or as ubiquitous as a security camera at a retail outlet. They can be as sophisticated as a drone making a 3D map in real time, or as dumb as a sensor in a basement detecting moisture. How powerful is this paradigm? Consider the supervisory control and data acquisition (SCADA) systems of 30 years ago, compared with those of today which are networked globally with access to elastic computing, crowdsourced data, artificial intelligence algorithms, and banking systems. Geospatially accurate, location-based data are poised to become the transactional currency of GEOINT.
However, GoT is not yet the overarching paradigm it is destined to become as GEOINT shifts from an age of paper maps and a handful of billion-dollar satellites to a new age of geospatial big data. In 10 years, the nature of GEOINT will be drastically different, and GoT will be a major element of that shift. Recent market reports on geo-location intelligence estimate that by 2020, 20 to 30 billion devices — not counting computers, smartphones, and other mobile devices — would be connected to the internet and generating location data.
Transforming the Ubiquitous IoT into the GoT
Smartphones are the most obvious element of IoT. There are now more than a billion of them all over the planet, even in the least-developed countries. Smartphones are on most of the time, and they move through space and time as people eat, sleep, work, shop, and play. These devices are connected to a global telecommunications network that interoperates across time zones, languages, political systems, financial environments, and social contexts. The devices tend to have sensors to detect shock, light, temperature, barometric pressure, battery status, radio frequency (RF) signals, and, of course, location. Cameras on these phones are getting more sophisticated with each new release — more and more phones are capable of conveying 3D depth perception using stereoscopic camera technology. Smartphones run applications that enable them to access and control other smartphones (e.g., family trackers and enterprise management software), wearable electronic devices (e.g., fitness trackers), lighting and thermostats (e.g., smart home switches and products such as Nest), and even unmanned systems and satellites (e.g., apps such as Parrot’s FreeFlight and Orbit Logic’s SpyMeSat).
Transportation systems are another great example of the way in which IoT has become part of modern life. By its nature, transportation is about movement, so an instrumented transportation system knows when things are where and whether they are moving or otherwise changing. Transportation “things” are now widely interconnected under the rubric of intelligent transportation systems. The things include vehicles, traffic control signals, navigation systems, tollbooths, parking meters, and even cargo. For example, vehicle telemetry systems such as GM’s OnStar know when a vehicle has been in an accident, prompting an automatic call for emergency services. Transponders report legal weight, safety ratings, and credentials as trucks continue down the highway while legally bypassing weigh stations. Other transponders automatically debit an owner’s account as vehicles drive through tollbooths. In many cities, parking meters are now connected to the internet and able to take payments by credit card. Even traffic signals are networked, some now with cameras designed to report red light violations, and others that can be centrally controlled to adjust for real-time traffic conditions. Most shipping containers are now tracked with RFID devices or scanner codes. And, most obviously, on-board navigation systems are connected to the internet — in many cases with two-way communication, so the system can learn from the experience of each navigator in real time. It’s not just cars and trucks, either. Airplanes, ships, busses, and trains are increasingly connected.
IoT as an Economic Force
One common misconception is that IoT is new. As sensors become ubiquitous, the actual cost to produce each unit declines. Marry decreasing per unit cost with improved communications bandwidth needed to support the volume of data analysis and transactions, and the value to governments and businesses at all levels continues to grow. Back-end processing is increasingly powerful and cost-effective.
More than nine billion devices are connected to the internet, including computers, smartphones, tablets, and more, and this number is expected to rise dramatically within the next decade. These devices make measurements continuously, thereby creating a history of measurements that can be analyzed in space and time to detect motion or other change.
In the weather domain alone, The Weather Company, now owned by IBM, ingests more than 100 terabytes of third-party spatiotemporal data daily from more than 800 different data sources, including satellites, personal weather stations, smartphone pressure sensors, and more. As one of the largest IoT platforms in the world, forecasts are produced for 2.2 billion locations every 15 minutes.
People who work in aviation, energy, insurance, media, and government rely on weather information for data, technology platforms, and services to help improve decision-making and respond to weather’s impact on business.
Instrumented, Interconnected, and Intelligent
There are many challenges to unlocking the value in IoT. Building the framework to support IoT infrastructure can be quite complex. In ensuring a robust infrastructure, one must support a world that is increasingly instrumented, interconnected, and intelligent. To address these IoT challenges, one must consider:
· The range or variety of devices: How to quickly connect a broad range of new and legacy devices.
· Awareness at scale: How to capture big data from devices at scale without stressing networks.
· Real-time analytics: How to analyze in-flight data to predict, detect, optimize, and anticipate.
· Easy orchestration: How to rapidly wire devices together and create logic without programming.
· Enabling Access: How to expose and monetize information and services while maintaining privacy and security.
· Geolocation: What level of accuracy and precision is needed; and, fundamentally, how to fix a location, especially inside structures, underground, or beneath water.
Is IoT-Derived Intelligence Really GEOINT?
Is IoT really GEOINT? Not traditional GEOINT perhaps, but if not GEOINT, than what kind of INT? Each thing exists in physical space at a particular time and may be permanent or ephemeral, static, or in motion. To the extent that each of these things collects and shares time- and location-stamped data about its environment or activity, IoT provides an unprecedented rich source of geospatial information. The volume of data is mind-boggling, and the data is constantly changing in response to natural and artificial activity. Imagine the possibilities for discerning patterns of activities and relationships among actors. Imagine the power to visualize connections between things and people as they change over time, and for simulating future possibilities.
This is not a future world. It is the world of today that we are just beginning to appreciate in a new way. So why don’t we notice it? Like the integrated circuit chip, IoT is embedded in our infrastructure and so tightly integrated with our daily lives that we take it for granted. And that is exactly why GoT can be such a powerful paradigm for making sense of all-source intelligence. Moreover, the physical nature of GoT makes intelligence information all the more actionable. Each thing exists in time and space, so each sensor measures uniquely from a particular location at a particular time. Time- and location-stamped data in motion forms a basis for historical analysis, state description, and predictive analytics, which can then be used to take action.
Trial Definition for GoT
What is the GEOINT of Things? Certainly, it is a new concept awaiting a definition, so here is a proposed definition: GoT is the intersection of rapidly evolving, interdependent, and widely available knowledge environments and technologies accounting for integrating, leveraging, and describing location, time, and relationships between humans, objects, activities, and their physical and terrestrial surroundings enabled by the networked age. In this paradigm:
· GoT impacts everyone connected to the IoT and even those who are not. Its varied effects (good or bad) also extend to the 4.2 billion without internet access due to decisions made and actions taken by governments, industry, non-government organizations, and private actors.
· Everyone is a sensor, to what degree depends on their level of connectedness, activity, and, exposure.
· Everything has the potential to be a sensor.
· Everything is somewhere, nothing is nowhere.
· Every measurement is made at a particular location and time.
· Because it is part of everyday life, location’s extraordinary value may easily be taken for granted, overlooked, or underappreciated, but nonetheless is relevant to daily individual and organizational activities.
· The IoT is dynamic — things are constantly in motion, and the data constantly changes with time.
Enablers of GoT
Another interesting feature of GoT is it inherently benefits from crowdsourcing of both data and analytics. Because things either make measurements, commit actions, or both, the greater the number of things and the more different kinds of things that participate in the data stream, the better. As the world becomes more connected, the data that record human and machine activity becomes increasingly diverse and powerful. Such large volumes of data can complicate the analysis, but commercial services that exploit particular aspects of the data use cloud platforms and crowdsourced analytics to make correlations for specific applications such as location-based advertising, shipping logistics, supply chain optimization, and even commodities trading.
To fully exploit GoT will require massive compute power and telecommunications pipes because the number of interconnected things is already in the billions and headed toward the trillions. Each of these things may be capable of making multiple measurements and taking multiple actions per minute. Many of these things will be moving, so historic locations, patterns of motion, and physical proximity of things will need to be stored and analyzed, leading to predictive models of future locations and interactions. It is easy to see how quickly the volume and velocity of data will grow over time. Imagine the current challenge of maintaining signals intelligence (SIGINT) capability, and then multiply the volume of data by trillions. Today’s SIGINT challenges will seem small compared with tomorrow’s GoT challenges.
Artificial Intelligence Meets GoT
Accompanying the revolution in location-based intelligence from IoT is a growing use and acceptance of artificial intelligence (AI) both in the fields of machine learning and computer vision. In a recent New York Times article by Quentin Hardy, major commercial software companies such as GE, Oracle, salesforce.com, and many others are investing in AI to unlock the hidden value of their data using technologies such as machine and deep learning. The traditional GEOINT applications of AI have largely focused on geospatial data production problems. With the advent of relatively inexpensive cloud computing resources, more AI applications are being developed to detect unusual patterns in data that more often than not have a location component. In an era of big data that includes mountains of Earth observation imagery, the shifting focus in the GEOINT Community is on the detection and visualization of patterns in geospatial data rather than extracting features such as the exact rooflines of a house. It’s great to have the building footprints, but the real money and intelligence is in the detection and anticipation of patterns and behaviors.
To meet these GoT-induced challenges, global GEOINT practitioners must adapt. The production of geospatial products (e.g., imagery composited from sensors that are part of the IoT) will be automated by necessity because no human will be able to keep up with the volume and velocity of data. Preliminary analyses will similarly be automated by necessity. Therefore, the workforce will need: the expertise to capture and organize the torrents of data from a wide range of time and location-enabled sensors; the domain expertise to recognize the difference between information and mere data; and the skills to visualize and form information products in a compelling way that enable decision-makers to quickly see what is important amidst a data sea of complexity, detail, and unimaginable quantities. The evolution of IoT into true GoT is already underway. It is up to the GEOINT Community to build into its existing business practices and academic curricula the methodologies and opportunities for current GEOINT analysts to work with and experiment with IoT data.
