The Conceptual Modelling of Dynamic Teams for Autonomous Systems

User and System Stories: an agile approach for managing requirements in AOSE

“…JACK [34], SARL [29], Jadex [28], and Jason [3]. We use SARL in our work, since it is what is currently used in our industry project and the TDF design artifacts are readily implemented in SARL (recent work [18] also used this combination).…”

User and System Stories: an agile approach for managing requirements in AOSE

“…TDF [15,16] is the successor to Prometheus [25], a mature and popular AOSE methodology. A pilot study has shown that TDF significantly improves comprehension of behaviour models, compared to UML [18]. The methodology follows 3 phases in an iterative manner: System specification where the system-level artifacts are identified, namely goals, scenarios, percepts, actions, data, actors and roles; System architecture where the internals of the system are specified, namely the agents that enact the different roles and the interactions between them; and Detailed design that defines the internals of the agents, namely plan diagrams, tactics and internal messages/sub-goals.…”

A Framework for Engineering Human/Agent Teaming Systems

“…We suggested above that humans naturally conceptualise a team as an intentional entity with joint mental state. We further argue that humans naturally conceptualise team cognition as being analogous to that of intentional individuals, namely humans, and for this reason we adopted the BDI paradigm to represent and present team behaviour in human/agent teams, and so used TDF-T (Evertsz, Thangarajah, and Papasimeon 2017) which is a BDI-based team modelling formalism, although other formalisms with similar teaming concepts could have potentially been used.…”

“…Based on earlier evidence that TDF-T is easy for people to understand (Evertsz, Thangarajah, and Papasimeon 2017), our hypothesis was that TDF-T diagrams would be an effective means of presenting important team-level information to the human so they can become familiar with the models offline. The guidelines we presented in the previous section foster a clearer representation, and this offline information was one aspect of our evaluation that we present ahead.…”

“…We have focused on using diagrammatic methods of representing and presenting the team's structure and decision making so that the human can see a visual overview of their role in the team, what goals they should be tackling at any given time, and the extent to which their activities need to synchronise with teammates. For the representation, we investigated the application of TDF-T (Evertsz, Thangarajah, and Papasimeon 2017) to the wider problem of engineering human/agent teams. We explored how to contextualise the TDF-T design diagrams by determining which aspects to instantiate and present to the human at runtime.…”