Omron Corp. has developed a machine automation controller equipped with a machine learning (artificial intelligence or AI) algorithm. The company claims the controller is the first of its kind, achieving real-time integration between the programmable logic controller function, which controls production lines and equipment changing in microseconds on factory floors and the AI processing function.

Omron said the introduction of AI and IoT on manufacturing floors allows manufacturers to contain the impact of skilled worker shortages and surging labour costs, while simultaneously increasing equipment utilisation and achieving stable production of quality products. To leverage data at manufacturing sites where control is performed in microseconds, high-speed and high-precision data gathering, such as position, vibration and temperature data, and precisely associating them with time data is required.

The controller uses Omron's sensors to monitor equipment and process status and predicts unusual movements of machinery based on the causal model that the built-in AI has been made to learn.

Omron claims it incorporates the knowhow of skilled workers into the controls of equipment and processes. This is accomplished by utilising a broad range of factory automation (FA) equipment that enables IoT-capable automated production or implements optimal AI algorithms in such equipment. Developed under this concept, the AI-equipped controller is meant to immediately detect signs of equipment irregularity.

The machine automation controller's AI algorithms allow it to learn the repeated movements of the equipment from precise sensing data, which enables feedback to status monitoring and real-time control of machines.

Here's how the company described the use of AI by the machine automation controller:

  1. Gathers sensing data and output data for motors from equipment and processes it chronologically in real-time.
  2. Generates characteristic quantities at regular or irregular times based on chronological data in real-time.
  3. Accumulates characteristic quantity data and generates model data of learning machine after causal analysis.
  4. Sends feedback to condition monitoring and controls based on such model data in real-time.

Sample shipments of the controller to some of Omron customers began in 2016 and that enabled the company to conduct demonstrations at its own and customers' factories, which is further enabling Omron to enhance knowledge of the cause-and-effect relationship involved in irregularities prior to the scheduled launch in 2018.