The global site of the UK's leading magazine for automation, motion engineering and power transmission
28 March, 2024

LinkedIn
Twitter
Twitter link

AI analyses workers' actions to improve productivity

18 October, 2019

Mitsubishi Electric has developed a technology that uses AI (artificial intelligence) to analyse the actions of workers on production lines and uses this information to improve their performance, thus boosting productivity.

Despite the increasing use of automation and robotics in factories, manual operations are still core processes in many factories, so improving work procedures remains a key to raising productivity, Mitsubishi says. At present, most manufacturers rely on supervisors monitoring line workers to observie them and detect any errors. But this is time-consuming and cannot usually be carried out on a regular basis.

In addition, workers often move too quickly for their actions to be observed reliably by human supervisors, making it difficult to obtain consistent and useful results.

Mitsubishi’s new technology, called Kotsumon, uses AI to extract data from standard videos of people at work on production lines, evaluating their activities and detecting any mistakes they make. The company says that manufacturers could use the results to improve productivity on their lines.

Mitsubishi predicts that its AI-based technology for analysing worker performance will boost productivity and save time

The system is claimed to identify motions with a 90% accuracy, without needing to adopt measures such as attaching sensors to the  workers. Mitsubishi says that an analysis that would take a human supervisor 30 days to observe and complete, can be achieved by the AI system in just three days.

The company, which developed the technology in collaboration with Keio University, plans to evaluate the technique on its own production lines before releasing it commercially, during the fiscal year ending March 2021 “or thereafter”.




Magazine
  • To view a digital copy of the latest issue of Drives & Controls, click here.

    To visit the digital library of past issues, click here

    To subscribe to the magazine, click here

     

Poll

"Do you think that robots create or destroy jobs?"

Newsletter
Newsletter

Events

Most Read Articles