In the public domain, the Internet of Things (IoT) is often spoken of in contexts of smart homes, wearable technology and connected toys. But IoT has far wider applications than that, from inventory management to shipping sensor systems. By 2025, the technologies making up IoT will continue to advance and be even more deeply embedded into the fabric of today’s society. For industries, this will mean not only increased understanding of the systems being measured and analyzed, but increased ability to remotely operate or automate those systems.
IoT is a key element of compositional architectures like digital twins, smart cities, “living” skins on airplanes and other vehicles, as well as more traditional applications such as oil wells. Sensors are also being deployed in such diverse applications as precision agriculture, medicine, and logistics. In its most basic application, the use of sensors provides insight on the status of a physical asset. Over time, however, the dramatic increase of types sensors and the quality of information generated has created a need for edge computing capabilities.
Always-on connectivity for homes and ships alike by 2030Today, simple sensing is being implemented across all industries, delivering a variety of data points to businesses. For example, the oil and gas industry is using highly instrumented hardware to measure the performance of the system and provide the organization with the data necessary to create predictive maintenance algorithms. The dramatic shift in the next ten years will be both in the utilization of these insights in the automation of various physical mechanisms and in the creation of new sensors and sensor arrays, as well as sensors integrated into materials to give us a richer understanding of our physical world.
IoT is a key enabler of digital twins in that it provides the insights necessary to understand the health of a physical asset in use. However, these solutions are dependent on high quality data, and schematics of the physical asset being up-to-date. As a result, there are challenges in keeping data in sync with the representation of the asset. Furthermore, many vendors and original equipment manufacturers are hesitant to open the data generated from their particular system to the larger digital twin, creating blind spots of visibility in the overall health of the asset. Whether the trend to seal off data continues or if it is opened up to third parties will significantly impact the value production from sensor systems.
By moving the processing of data generated by IoT sensor arrays to the edge - in other words, physically closer to the asset that they monitor -, latency can be reduced dramatically, providing real time responsiveness to machine learning and algorithmic capabilities. These capabilities can also be lent to sometimes-connected states, or states where always-on connectivity isn’t economically viable, such as on vessels that use satellites for connectivity. Advances in satellite-based communication are expected to dramatically decrease the costs there as well.
While organizations today are already leveraging IoT for tasks such as predictive maintenance and operational efficiency gains, the growth areas will come in leveraging IoT in conjunction with other emergent technologies such as edge computing and AR/VR/MR. The timeframe for adoption in an organization will vary widely based on factors like data maturity, finding the best application for physical sensors, and clear strategies for how to apply the data collected from sensors.
By 2025, more sophisticated multisensory arrays in smaller physical form factors will emerge, as well as sensors being woven into physical forms creating “living” materials. For example, aerospace is trialling 3D printing of wings using carbon fibre with sensors woven into the process.
As IoT already has a strong foothold in a variety of industries, growth will continue in the ongoing instrumentation of the physical world. There are some industries that have been hesitant to embrace these capabilities but are quickly seeing the value and are rapidly ramping up their capabilities in this space. As with many emergent technologies, adoption continues to be limited by the human resources necessary to design, implement, and operate these technologies. Another challenge is a lack of consistency or standardization in the data formats being generated, leading to backlogs in data science workloads requiring data quality assurance.
Another point of uncertainty facing organizations is what to do with the volume of data once it is aggregated. With the goal of collecting data to improve automation in physical systems, more and more organizations are adopting strategies around using machine learning to train systems based on data in flight (hot data), and then keeping a sample or subset of data over time (warm or cold data).
Therefore, data quality assurance and system verifications will be essential for a healthy IoT system as the associated technologies ramp up development and data generation in the coming decade.
ContributorsMain author: Chris Pelsor
Editor: Tiffany Hildre