Issue link: https://resources.mouser.com/i/1442808
7 The Convergence of Enabling Technologies The IIoT is not just another new technology. Instead, it is a collection of multiple technologies that enable high-worth applications. These enabling technologies have matured over the years to a point where their capabilities can be combined to design cost-effective products and services on a larger scale. In general, any IIoT use case requires the ability to collect, monitor, exchange, analyze, and deliver insights to trigger an action. A few key technologies that enable these functions will get discussed. Sensing Technology The past decade has seen significant advances in sensor technology, comparable to a renaissance of sorts. With the sliding cost and miniaturization of micro-electromechanical systems (MEMS) technology, it is now much easier to integrate sensors into products. Sensors are now widely utilized in automotive systems, industrial control, healthcare, oil exploration, climate monitoring, robotics, and in many other applications: For instance, a smartphone boasts of at least eight to ten sensors. In a wind turbine, which often gets mounted on a high tower (~200m) with long rotor blades (~80m), the most critical components are the sensors. These sensors, though only a few centimeters in dimension, play an outsized role in predicting and preventing failures in the face of extraordinary stresses, vibration, and various other hazards. A sensor's data provides operational insights. However, in the IoT context, sensor data is more effective when it can be fused to provide contextual information. The sensor fusion technology enables a microcontroller to merge the data from multiple sensors to create meaningful contextual awareness. This technology is used to design robotics and other autonomous machines. Communications and Computing The sensor data is useful only when the information is capable of being safely communicated to an analytics platform. While machine-to-machine (M2M) communication has already been present in industrial use cases, recent advances in radio communications and cloud computing have received the credit as being the key drivers in the progression of the IIoT. Due to privacy concerns, enterprises have restricted the use of cloud technology for many years. But with increased security and computing horsepower, the cloud is becoming an undeniable component of any IIoT use case. Multi-access edge computing (MEC) is a parallel development that can be utilized to offload cloud processing to platforms located closer to the sensor nodes. MEC allows the cloud providers to cost-effectively provision and manage millions of IoT devices, compute and analyze sensor data near the source, and host applications to generate actionable insights. Open communication protocols such as the Ethernet, 3G/4G, LoRaWAN, and Zigbee can now be integrated into IIoT product design and managed through handheld mobile devices. Machine Learning and Artificial Intelligence For autonomous decisions, sensing and connectivity are not enough. We need intelligence, as well. Cognitive technologies have been foundational to design intelligent, connected products, whether it is an industrial robot, an autonomous vehicle, or an enterprise wizard-like IBM Watson. Machine learning and artificial intelligence (AI) are two mainstream cognitive disciplines to design products that can learn, think, adapt to new environments, and make decisions with hardly any human intervention. Some of the critical capabilities and technologies in this context are: Perception Perception is enabled by sensors that allow for the detection of ambient parameters, like distance, proximity, force, torque, velocity, sound, vibration, volume, weight, surface finish, texture, chemical composition, temperature, electrical current, and magnetism. Machine vision is another technology to detect and identify objects, spaces, scenes, orientations, and locations of objects in the visible, infrared (IR), and ultraviolet (UV) wavelengths. The convergence of these perception technologies improves accuracy, throughput, collision avoidance, and situational awareness. Voice recognition and motion tracking technologies are also useful to interact with the environment and with human operators. Data Fusion and Processing It is essential to process data fused from multiple sensors and actuators to optimize and improve accuracy. Sensor-fusion algorithms enable the effective integration of data from different types of sensors. This data may come from various locations within the workspace or may be sensed at different times to make informed and accurate deductions about the environment and situation. Deep Learning, Neural Networks, and Natural Language Processing Sensor-enabled machines and devices generate massive data. The rapid improvements in parallel processing power, AI, and machine learning algorithms, as well as the availability of large volumes of IoT data sets to feed these algorithms, are together enabling the design of intelligent, cognitive IIoT systems. Advancements in deep learning techniques, neural networks, fuzzy logic, and natural language processing (NLP) are opening newer cognitive capabilities, enabling systems which can understand, learn, predict, adapt, and potentially operate autonomously rather than execute predefined instructions. Actuation In the design of robot-arms and other robotic applications, expensive, high-force actuation devices or motor drives are commonly employed. However, for designing lighter and nimbler robotics, it is now possible to use less-expensive drive mechanisms or motion sensors that rely on force or current feedback to compensate for potential errors in position accuracy. IIoT Applications and Trends The convergence of these enabling technologies is fueling several applications and trends in the industrial sectors.