What makes Industrial Machines Smarter and Intelligent in the Future?

Veera Manikandan Raju, CTO- Software & Systems, Connected MCU, Texas Instruments India | Tuesday, 28 November 2017, 12:22 IST

Artificial Intel­ligence has emerged as one of the prime dis­ruptive forces cutting across different indus­tries –well poised to influence the next industrial revolution. Hydraulic energy, Electrical Energy and Com­puter Aided Design were the primary forces of influencing past three in­dustrial revolutions – now the time has come for AI driven cyber physi­cal systems to change the landscape of manufacturing in industries. Ma­chines and processes communicat­ing with each other without human intervention is the vision of fourth industrial revolution.

What is making this revolution to be closer to reality? Artificial Intelligence for Industry 4.0 is closer to reality now because of rapid advancements of few technologies fueling this transformation. Examples of such technological advancements are - deeper intelligence at sensing nodes and actuators, deep learning advancements in machine vision with embedded processors, 3D printing, distributed security using block-chain, ubiquitous connectivity in industrial communications with low latency and high data rates and enhanced storage capacity per 2mm of silicon at lowest power.

Thanks to rapid progress in innovation in embedded sensing and MEMS – a wide variety of rugged industrial sensing using fusion of multiple sensors like magnetics, ultrasonic and optical sensors is a reality today. We would see more and deeper intelligence at the sensing nodes evolving in coming years to ensure that the data available at the edge nodes is good enough to take the decisions independently without reaching the cloud. Think of a recipe driven manufacturing in pharma industry – a bottle of syrup to be manufactured – will go through a conveyor belt and will clearly know at what level the liquids needs to be mixed and what kind of labeling needs to be added in the bottle without any manual intervention. Think of the raw materials needed for the syrup that are automatically updated in the backend systems for robots to re-place and deliver the new liquids through efficient and predictive supply chain management. Embedded sensing when enabled with smart analog frontends and tightly coupled embedded processing are well poised for correcting non-linearity and imperfections in measurement and positioning systems.

Control frequency and sampling rate of industrial applications require very low latency communication protocols – like Ethernet switches at the industrial networks need few micro seconds latency in controlling the actuators like stepper motors. A hybrid utilization of wired and wireless connectivity enables a highly distributed network of intelligent machines and parts on the factory floor or in warehouses. Assigning IP addresses directly into machines and transducers is highly desirable in intelligent factories. With advancements of 5G networks, technologies like narrow band internet of things (NB-IoT) would be able to uniquely identify each equipment using an IP address in the factory floor digitally and would be able to communicate independently. Robotic actuators with flexible sensing and arms would make the ware-house automation to next level. E-commerce fulfillment warehouses are under tremendous pressure to meet increasing demands for fast, accurate order fulfillment despite having huge amount of labor-force in countries like India. We see increased number of startups in this area of Automation to optimize the inventory and delivery costs for e-commerce giants.

Biggest resistance and threat to the industry 4.0 revolution is the data privacy of the manufacturing houses and the security of their internal data. Blockchain as a distributed security technology is well poised to enable the distributed asset management with highest security for factory floors. Embedded security capabilities at each of the sensing nodes with secure re-provisioning capabilities, will make this barrier go away soon and industries would adapt wireless networks in factory automation segments.

Transferring high data rate information and wideband signaling for measurement and control across isolation barriers is increasingly important in industrial systems. Robustness against high voltage surges is indispensable for industrial interfaces as the cyber physical systems are highly integrated with high voltage and low voltage circuitry.

We see convergence of many technologies enabling the artificial intelligence to make the machines smarter. Microcontrollers made for deeper intelligence, embedded processors with highest compute power, intelligent sensing nodes with sensor fusion, integrated non-volatile storage in embedded processing at lowest power, high precision and ultra-low latency communication systems to control the robotic actuators, distributed security capabilities with blockchain technology at the factory equipment, ubiquitous connectivity with unique digital identity are the few that semiconductor industry is driving at rapid pace to make such a transformation reality soon. Think of machines directly self-repairing itself and calls upon the operator for human intervention only when needed – what a way to avoid night shift operations by humans in factories.

Overall this era will see a fusion of digital, physical and biological systems together to form future industrial machines – which will drive real-time optimization of the entire manufacturing flow with very minimal human intervention, with the complete usage of fossil fuels and produces zero industrial wastes ensuring green and safe environment for us to live peacefully. Smarter Industrial machines are not a distant dream anymore as we see the transformations and adaptions in the industry happening faster. Will the Internet of bio-nano Things (IOBNT) add a biological advancement to the machines for human like flexibilities and movements in factories? Time will answer.