Enhancing Performance in Human–Robot Collaboration: A Modular Architecture for Task Scheduling and Safe Trajectory Planning

The integration of robots into shared workspaces alongside humans is the basis of Human-Robot Collaboration (HRC). This field of research has changed the paradigm of the industrial context, making HRC of pivotal importance for both researchers and the industry. In this context, a suitable task scheduling and trajectory planning strategy are crucial to achieve good performances and create a synergy between the two actors. Indeed, the task scheduling should be able to optimally distribute the tasks between the actors and recover from possible failures, i.e. by rescheduling the tasks. The trajectory planning strategy must comply with the safety standards that impose a reduction of velocity based on human behaviour. To this end, the monitoring system must also be safe-certified; otherwise, safety cannot be guaranteed. This paper proposes a novel architecture that integrates a dynamic task scheduling module with a dynamic trajectory planning module that explicitly considers ISO/TS 15066. For this purpose, the framework exploits a secure and certified monitoring system capable of tracking the human operator even in case of occlusions. The overall platform has been extensively validated both in a real and complex industrial scenario within the context of the ROSSINI EU project, where a dual-arm mobile robot collaborates with a human operator in an automatic machine-tending operation, and in a mock-up scenario. Note to Practitioners—This paper focuses on the implementation of a modular and integrated architecture to handle the challenges associated with effective task scheduling and trajectory planning, with the aim of improving the performances of collaborative robots in real HRC scenarios. The proposed methodology is structured into three distinct layers. Firstly, the safety layer monitors in real-time the human operator and provides the other layers with the maximum admissible robot speed. Secondly, the scheduling layer distributes tasks between humans and robots efficiently, recovering from possible failures and rescheduling the tasks, if necessary. Lastly, the flexible execution layer plans collision-free paths and maximizes the robot speed while explicitly considering ISO/TS 15066 guidelines. Thus, it ensures that the robot behaviour is compliant with the safety standards. Although the overall architecture is fully integrated, each layer is designed to function independently. Practitioners have the flexibility to utilize each layer as a standalone module if desired. Moreover, the architecture has been validated in a real complex industrial scenario, proving its practical applicability, robustness, and effectiveness.