EDDE: An Event-Driven Data Exchange to Accurately Introspect Cobot Applications

Commercialized collaborative robots (cobots) are typically programmed in proprietary languages and employ interfaces that gather data at predefined frequencies, such as the Real Time Data Exchange offered by Universal Robots (UR). This approach challenges the observability of program execution during runtime. To address this issue, the paper proposes applying software architectural knowledge from the introspection of Cyber-Physical Systems (CPSs) to the robotics domain. The proposed solution is to improve the classical protocol interfaces with an Event-Driven Architecture, enabling the acquisition of program events in relation to state variables and thus, richer information concerning runtime. The proposed architecture, called the Event-Driven Data Exchange, is implemented and evaluated on a UR e-Series cobot in a practical study. The study highlights the significance of implementing software architectural knowledge in CPSs, such as cobots to extend observability and data availability.