Today, automation has already transformed industries in which complexity and performance demands must meet the challenges of scarce resources, narrower profit margins, and expanding product volume. And now, thanks to the rapid growth of innovative technologies, such as 5G, AI, and IoT, industries are inching ever closer to fully autonomous technologies.
While current automation involves programming a system to perform specific tasks, autonomous systems are programmed to perform automated tasks, accommodate for variation, and self-correct or self-learn with little or no human intervention. Automated systems are especially invaluable in the case of complex processes since they can most efficiently address anomalous conditions with rapid, real-time sensing of the departure from the normal operations and can address them quickly. In the majority of cases, such a process is vastly superior to the management of the process by humans in terms of its effectiveness and operating costs.
For IIoT applications, data is being generated in greater volumes than ever before, and the plant IT/OT infrastructure is evolving rapidly with the latest generation of edge devices, actuators, and controllers, such as DCSs, PLCs, and PACs. Process data collected through networked sensors continuously at the Device Edge is the basis for automation and control of production assets and industrial processes. “Device Edge” of the network is where sensors and other devices generate data and connect directly to communication devices, such as gateways and hubs, prior to any WAN connection.
Local processing of the data is preferred in many cases since transmission of large volumes of raw data to the Cloud through wired or wireless public networks is a large recurring expense. If data ownership is a key consideration, data sent to the Cloud is also subject to data breaches depending on the measures taken to protect it. The optimum design is one where the IT infrastructure that collects the data throughout the plant can also aggregate, pre-process, filter, and relay the processed data and run the analytics tools to provide instructions that are then fed back through the network to actuators or controllers to maintain the same operations or modify the processes. This effectively provides for a continuous feedback loop that adjusts automatically, typically supported by machine learning and AI, and makes human interventions unnecessary in most cases, especially for highly complex processes. The response time through this feedback loop for a vast majority of the solutions that depend on process automation is critical to avoid damage to the equipment and minimize various risks, including worker safety and costs, through corrective actions.
The fastest response times for process automation are typically achieved through local processing of data prior to the WAN connections. However, the low latency connections achieved through most 5G implementations provides for a multitude of opportunities to employ any combination of public and private 5G networks with local processing to pre-process the raw data to minimize the bandwidth requirements for 5G connections while achieving a large degree of data privacy by maintaining the “data lakes” and “data warehouses” local.
This leap from programming to performing tasks autonomously is being accelerated as more hyper-converged edge computing with 5G services becomes available, bringing with it next-level speeds compared to 4G. Though current 4G has a latency of 50 ms, it’s not low enough to account for split-second responses needed in autonomous systems. Latency is usually estimated as a round trip delay, calculating how long it will take one system module to send a message and then receive confirmation of receipt from another system module before it is able to send more information. Low delays achieved by the development of 5G-based mobile networks open the way to radically new experiences and opportunities, including factory robots and other applications for which a quick response is not optional at all but a strong prerequisite.
“With the rapid convergence of 5G, the evolution of AI, the enormous growth in IoT and Industrial IoT, and the movement of the cloud closer to the edge, we are at the threshold of unlocking high value, real-time autonomous capabilities at the edge,” said Allen Salmasi, founder, and CEO of Veea, a pioneer in edge connectivity, computing, and security.
Salmasi also explained that ecosystems are flourishing, in many ways inspired by access to faster and more resilient 5G broadband capacity. “Autonomous, connected solutions are not easy to build, but they are easier to implement, manage and scale more than ever before, and represent the next major inflection point for edge computing and Industry 4.0, across a new and open frontier for business innovation and reinvention.”
Definitions of autonomous systems vary, but the growth of real-time control systems is indisputable. In 2018, Siemens shared that 2005 spending on robotic systems was $11 billion and that by 2025 it is expected to reach $67 billion.
Moreover, IIoT has become a common denominator for OPC Unified Architecture (UA), which is a family of technologies intended to exchange information between different industrial control systems on different control layers in a standardized fashion. OPC UA provides technologies to connect field control devices with higher control layers. OPC UA has been gaining traction as a standard for connecting the plant floor to the enterprise. The OPC Foundation pursues interoperability in industrial automation by creating and maintaining open specifications that standardize the communication of acquired process data, alarm and event records, historical data, and batch data to multi-vendor enterprise systems and between production devices.
“Beyond this amazing growth story is the trend that instead of merely performing repetitive tasks, robots are heading for incremental autonomy,” the company said. “Combine this with Artificial Intelligence, a space Siemens has been active in for more than 30 years and holds some 50 patents in machine learning processes alone – and we can clearly see a trend that is set to transform not only production and logistics but business models and user behavior.”
“Autonomous Systems at the edge are similar to robotic process automation, using the attributes of ubiquitous intelligence, hybrid edge-cloud computing with 5G network connections, machine learning, AI and autonomous infrastructure with the potential to transform our world again in Smart Cities, Smart Campuses, Smart Homes, Businesses, Transportation Systems, Factories and more,” Salmasi said. “IoT and Industrial IoT is key to expanding our vision for systems that today can run based in part on sensor data and code, working locally at the edge using mesh networking, and working in a distributed fashion when data is shipped to the cloud for a view of multiple edges, for example, automotive manufacturing plants spread out across the globe.”
Applications, and the associated ROIs, across Automated Systems, are emerging rapidly. The business logic and vision have been in place for years, as enterprises seek to reduce costs, improve safety, ensure accuracy and quality, and deliver more competitive products and solutions to the market, Salmasi explained.
“As intelligence becomes more ubiquitous, and as cloud computing models and technologies make their way to the edge, and as edge infrastructure, including at the “Device Edge” of the network, where sensors and various devices connect directly to Smart Computing Hubs for local processing prior to any WAN connection for the fastest response times, the possibilities are limitless. That’s why Veea has been investing in and building what is now required to make Automated Systems not only work but work at the highest level of efficiency and security, minimal maintenance, and most competitive total cost of ownership for such use cases.”
While some early automated systems function with 4G or wired ISP WAN connections, the Autonomous Edge relying on 5G network connections will make it possible to run universal applications in different geographic locations and provide for distributed automation with federated machine learning to improve the process automation but also a means to ship data from edge deployments into ERP systems which can be viewed by a single administrator or analysts on a single pane of glass.
“To fully support powerful autonomous systems that function predictably and efficiently at the edge, we have created hardware, software, and a platform as a service that supports the technical requirements, including supporting standards-based wired and wireless protocols, while addressing the need to orchestrate resources deployed at the far edge of the network.”
Because of the real-time benefits of autonomous systems, a variety of industries are already turning to innovative technology. For instance, the global autonomous construction market is predicted to rise by 2.5 times from 2019 to 2027, becoming a $12 billion market with a more than 15% growth rate over a similar period. At the same time, the autonomous navigation market is estimated at USD 2.2 billion in 2018 and is projected to reach USD 13.5 billion by 2030, at a CAGR of 16.19% from 2018 to 2030.
We are at the beginning of a revolution, where machines are called upon to progressively replace humans in their capacity for situation awareness and adaptive decision making. For companies in the industries cited above, as well as a variety of others, it’s now just a matter of when they will adopt AI and autonomous systems, not if. Ultimately, the extent to which we will use and benefit from autonomous systems will depend on how much we trust them.