To transform your organization’s physical field operations, you should have a strategy to (1) establish a Common Operating Picture by aggregating your siloed IoT systems, (2) contextualize your IoT data so they are AI/ML ready.

The Internet of Things (IoT) is vast and continues to expand at an astounding rate. It’s estimated that by 2025, 21,500,000,000 sensors and actuators will be embedded in real-world objects. All of these objects will be connected to the Internet and generating data at a high velocity continuously. The majority of these objects will be Industrial IoT devices used by industries, not consumers, with experts predicting a mind-boggling 847 zettabytes of IoT data generated annually by 2021.

As the wireless communication and sensor technology continues to improve and become exponentially less expensive, this unprecedented growth will continue.

The IoT represents the next evolution of the Internet, taking a huge leap in its ability to gather, analyze, and distribute data. It represents a significant business opportunity, providing organizations with strategic insights and operational efficiencies in real time. But harvesting the benefits of a sensor-rich world presents many data management challenges.

IoT Data is Everywhere

IoT data are generated by sensors and actuators that are embedded in real-world objects. As a result, IoT data is everywhere. It’s derived from multiple disparate systems and can cross a variety of activities that incorporate aspects such as location, time, conditions and movement.

Once collected, it is typically stored in multiple information silos across an organization. Each silo of data will differ. Collection cycles will not align and the data fields themselves may vary by sensor. These inconsistencies must be resolved before the data can be properly aggregated and used.

Most of the IoT systems are siloed systems built for a single application. Organizations know that they need to combine them together and run AI/ML models so that they can be more efficient and competitive. But they cannot, because these systems are siloed and cannot be easily combined.

Sometimes, necessary data points can be overlooked altogether. I have seen so many IoT systems fail to capture necessary information, such as time zone, unit of measurement, and geographical coordinate system information. These oversights render the data meaningless since it is impossible to determine accurate times, locations, and even sensor readings around the different events that occur.

The other common mistake I have seen is assigning the location of sensors to the location of observations. This is an important distinction. The data generated from drone or camera systems represents where the cameras are looking, not the location of the drones or cameras themselves.

With all of this said, disparate data sources should not be construed as a bad thing. In fact, the more data sources the better. If done properly, this variance will provide more accurate insights for the business. However, in order to achieve these results, the data must be properly prepared.

IoT data is better when they are combined

In order to compete in the sensor-rich world, companies must shift from disparate IoT silos consisting of closely related sensors, to a system that links together an array of different sensing systems.

For example, at an oil well site, oil is pumped into a production tank. When the tank is full, the liquid is transferred to trucks for transport and distribution. The elusive sweet spot — and how IoT data can help — is enabling a “just-in-time” operation, i.e., aligning the exact moment the tank is filled with the arrival of the liquid hauling truck, as well as the availability of the personnel to oversee the transfer. Three separate work streams, each incorporating endless scenarios that can (and will) impact the outcome.

Situational awareness of disparate field assets is critical in order to enable a just-in-time operation. That means there is no wait, no queue, no downtime, and no unnecessary site visit.

The calibration of all these moving bits is further complicated by additional factors. Bad weather, various work permits restrictions, and maintenance issues could delay the liquid hauling operation. The tank may be filling slower or faster than anticipated as the production rate changes constantly. The variables are endless.

Without aggregating and harmonizing IoT data from different systems to deliver these insights, the oil and gas company is unable to learn from the rich data they’ve already spent millions of dollars collecting. This means they lack the insights to properly automate schedules and tasks in order to optimize efficiencies.

Instead, tanks are full and idle as they wait to be emptied into trucks that have not yet arrived by work crews who are not properly prepared to accommodate the transfer. Expensive inefficiencies that can be avoided when IoT data is properly utilized.

Unleash the value of your IoT Data by making them AI-ready

An effective data management system for IoT data must utilize a repeatable and reusable model to ensure consistency in how the data is cleaned, enriched, and aggregated.

I’ve spent the past 14+ years focused on these types of challenges. One outcome of this work is the creation of the Open Geospatial Consortium (OGC) SensorThings API, a set of international, military-grade standards for device-cloud communications and cloud-cloud communications. The SensorThings API provides an open and unified way to interconnect IoT devices, data, and applications over the Internet.

Context and Certainty

IoT data streams present complex issues related to data quality. Data is often missing or subject to noise and calibration effects. For example, IoT devices typically generate highly nonuniform observations in space and time. By using the OGC SensorThings standard as a universal framework, we’re able to apply quality control and context to IoT data in a consistent manner — an imperative for processes that rely on Artificial Intelligence (AI) and Machine Learning (ML).

At SensorUp, we use a proprietary repeatable data ingest process to perform this work quickly and efficiently. This includes the necessary gap-filling, re-gridding, and calibrating algorithms that operationalize on a scalable computing infrastructure. Once cleaned and enriched, the IoT data can be properly aggregated so it can start to deliver the insights and business intelligence that will truly benefit your business.

OGC SensorThings API Standard offers a future-proof and unified framework to aggregate IoT data from multiple systems and make them AI/ML ready.

What does this all mean for your business

The data exists, you already have it. Once properly prepared and optimized, you can transform these endless stores of IoT data into business intelligence that will drive meaningful results for your company. As a result, you will be able to perform just-in-time operations, with no wait times or waste, and zero safety incidents.

I am always excited to share my thoughts on AI, IoT and digital transformation in industries such as oil and gas, rail, mining, etc. If your company or event is interested in having me speak or run a workshop, please reach out to me at steve.liang@sensorup.com.

An IoT-enabled Common Operating Picture powered by SensorUp. More than 20 different IoT systems in real-time loaded, transformed, and aggregated by using the OGC SensorThings API standard, so that situational awareness can be established and actions can be triggered in an emergent situation.