The AI-equipped controller is said to achieve real-time integration between production line controls, and AI processing. Using inputs from sensors monitoring the status of equipment and processes, the controller applies a causal model, learned by the built-in AI, to predict unusual machinery behaviours and to control them safely before problems can occur. The AI algorithms allow the controller to learn about repetitive machine movements from the sensing data, and to monitor the status of machinery and control it in real-time.
Omron began sample shipments of the AI-based controller to selected customers last year, and has conducted demonstrations at its own and customers’ factories to predict and analyse equipment malfunctions and to enhance knowledge of the cause-and-effect relationships involved in these malfunctions. It plans to launch the AI controller commercially in 2018.
In 2015, Omron revealed plans to make all of its 100,000 varieties of FA equipment IoT-capable. It has since begun releasing IoT-based equipment, while at the same time acquiring expertise in AI-equipped mobile robots and other types of industrial robots, and offering such technologies globally. It recently released a temperature controller with built-in AI, which it claims is the first of its type, and is currently conducting trials at its own and customers’ factories in the Netherlands, China, and Japan to determine how different types of FA equipment, including sensors, can be equipped with AI algorithms and how such equipment can be made IoT-capable. Its aim is achieve manufacturing processes that “produce no defects and do not stop”, using AI-equipped, IoT-capable devices to monitor the status of equipment and processes and to assure product quality.
Omron says that recent years have seen a shift towards small-batch production and manufacturing at optimal locations around the world. It argues that these trends are driving the need for AI and the IoT on manufacturing floors to limit the effects of skills shortages and labour costs, while simultaneously increasing equipment utilisation and achieving stable production of quality products. To make use of data at manufacturing sites where control is performed in microseconds, high-speed gathering of sensor data is essential and this needs to be associated precisely with time data. The combined data can be analysed and used to predict possible machine errors and to prevent stoppages as well as product quality deteriorating.
At the same time, the rapid adoption of IoT has brought with it huge quantities of data, making it difficult to transmit all of the data to the cloud for analysis by AI algorithms, regardless of how much network bandwidths are extended. Another bottleneck in introducing IoT to manufacturing floors is the delayed responses caused by the bidirectional transmission of data generated by sensors, motors, and other devices which must be sent to and from the cloud. This makes it difficult to achieve the instantaneous response that is crucial at many manufacturing sites
By equipping automation equipment with AI algorithms that replace the knowledge and intuition of skilled engineers, Omron says it is working to achieve the “manufacturing floors of the future” in which “machines bring out human capabilities and creativity”.