Researchers from Swinburne have automated the processing of blood donation packs, handing the tedious (and super delicate) job over to robots.
To clarify, the robots aren’t taking any blood from humans, rather they’re taking over once the donation packs have been received.
The project is described by Swinburne as one to “automate the folding and centrifuge loading process using collaborative robots, vision systems, jigs and actuators”.
What does that mean, exactly? At a processing centre, whole blood donations must be separated into its cellular components via centrifugation. To do this, the blood pack must be folded in a particular way to ensure during the process there is no bacterial contamination (which Swinburne said also increases shelf life).
Why hand this task over to robots? Well, even with highly trained staff, there’s still risks and instances of human error.
“Damaged or torn packs not only lead to the loss of a precious donation, but also disrupt production and expose staff to potentially hazardous biological materials,” Swinburne explained.
“Even subtle non-conformities can occur and build over time, leading to quality deviations. On the other hand, such repetitive motions – sometimes hundreds a day – can cause ergonomic strain and injury for staff.”
It’s not that straightforward, however. Swinburne said folding is a highly complex procedure that is challenging to automate using robots. It said the blood packs are soft and “deformable” objects, which can lead to a significant variation in shape and geometry.
Basically, this makes it difficult for a robot or computer that is not suited to the myriad of geometries. The Swinburne team saw this near-impossible task as opportunity, however. They proved that taking the task and breaking it down into smaller steps, as well as building in design and engineering contingencies, meant that the final design could involve a combination of semi-automated, automated or assisted processes.
Following their success, the team is now extending the project to focus on improving the speed of the robot – which is currently slower than a human.