The collaborative screwdriving process presented in CIRP-CMS 2023

Elias Montini has presented the studies conducted to integrate a collaborative robot to a manually operated screwdriving work cell in the CIRP-CMS 2023 Conference at Cape Town, South Africa on 26.10.2023. The collaborative robot working together with a human operator speeds up the screwdriving process 17% in GHEPI Srl in Italia.

Collaborative robots

Unlike conventional robots, which have higher price tags and less flexibility, collaborative robots are much cheaper, easier to program, and capable of working with operators without the need of a protection cage. This ‘collaboration’ not only increases production efficiency, but also decreases the physical and cognitive stress of the workers and enhances job satisfaction.

The proposed framework

The paper presented in the conference aims to propose a framework to help practitioners and researchers in implementing human-aware collaborative robotics systems. The framework supported the development of a collaborative screwdriving application in which both the operator and the robot support and perceive each other to optimize process performance and worker well being.

The proposed framework is grounded on 3 pillars (H: Humanisation, S: Smartification and A&E: Automation & Equipment) and 7 building blocks (Human factors, Training, Ethics & Trustworthiness, Performance, Cobots, Human-Aware Digital twin and Reconfigurability)

The status before the experiment

The as-is scenario foresees a work cell equipped with a 500-tons injection moulding press producing heavy plastic components, which are assembled by a human operator performing the collection of the two moulded parts from a conveyor belt to a workbench, positioning and tightening 9 screws, flipping the assembled part, positioning another 9 screws and tightening them, and finally stacking the assembled part on a pallet.

No automation systems (e.g., cobots) are currently used in this work cell, with the operator handling such heavy parts and controlling the complex production process.

The status after the experiment

The to-be scenario changes the original workflow in two points: while the operator is carrying out ‘positioning’ step, the cobot starts the screwing operation in slots that are ready to be tightened and can be reached with a collision-free trajectory. Therefore the overall cycle time can be reduced by performing the positioning and the screwdriving tasks in parallel.

A critical human factor to monitor is the worker’s perceived level of fatigue to consider for tasks assignment. The human-aware Digital Twin is instantiated with a platform for building DTs that provides a reference data model to structure the relevant entities and information. The platform creates a digital representation of the worker and of the work cell, including the current perceived fatigue of the operator, operator’s hands position and the positions of the screws and their status (tightened or loose) detected by a camera.

Conclusions

The main objective of the pilot experiment was aiming to decrease the cycle time of assembling by 20%. The integration of the collaborative robot decreased the cycle time from 152 seconds to 126 seconds (17% improvement) besides lowering the physical stress of the operator, increasing the job satisfaction and providing a better workload balance.

The scientific paper

The paper presented in the conference can be accessed via the following link:

https://www.researchgate.net/publication/374978814_A_Framework_for_Human- aware_Collaborative_Robotics_Systems_Development

The lighthouse demonstrator

A smaller scale demonstrator of the whole study can be seen in the Mini-Factory of the SUPSI Sustainable Productions Systems Lab (SPS-Lab).