Robotic breakthrough made by Manufacturing Technology Centre

Latest breakthrough made by automation experts at The Manufacturing Technology Centre (MTC), means robotics can handle objects without the need for CAD data or designed grasps to enable faster deployment of robotic solutions.

“This project demonstrates MTC’s determination to adopt academic developments with potential to transform robotics in manufacturing. The transferability of this technique to pick new objects will allow MTC to quickly test with customer parts and advise on implementation strategies,” commented Mark Robson, Senior Research Engineer, Robotics, MTC.

The project

Bin picking, a common handling task in industries where a single object needs to be separated from an unstructured bulk input. 

With traditional methods, factories use CAD data alongside high cost sensors to identify individual parts accurately, as well as testing a range of pre-engineered grasps for feasibility. 

To reduce time and costs the MTC has demonstrated its ‘state of the art’ technique for bin picking. The technique uses a trained model to find the best position to place a vacuum cup, and can be trained with simulated data to reduce the need for labour intensive manual data collection and labelling.

“MTC Robotics Engineers are proving that design and development can be dematerialised in the creation of cost-efficient intelligent solutions to automate operations where predetermined programming is less viable,” added Dr Alejandra Matamoros, Technology Manager, MTC.

The outcome

The project technique was tested by MTC on a range of objects including metal components, fruit, and cosmetics containers. 

Trained with month Manually labelled data and simulated date, the technique performed well with 92% and 94% successful picking respectively.

“Performance of the model trained on purely simulated data showed that this approach is a good solution to reduce the burden of data gathering for specific use cases,” stated MTC, adding that “excellent performance on items not present in the training data showed that the method generalised well to any items.” 

For more information on manufacturing topics - please take a look at the latest edition of Manufacturing Global.

Follow us on LinkedIn and Twitter.

Image source