Plan Negotiation through Cooperativeness Characterisation in Human-Robot Teams

This paper presents a cognitive characterisation framework for human-robot plan negotiation (HRPN) that situates proposals within the cooperativeness space , a cognitive model that separates stakeholder-directed concern from agent valuation and behavioural factors. This framework is instantiated in a non-critical service-robot scenario and evaluated through a two-phase experimental study in the human-robot collaborative search (HRCS) domain: a first phase to elicit human valuation and a second phase to examine real-time plan negotiation dynamics. This exploratory study characterises, for the first time, plan negotiation in a human-robot dyad through the lens of individual and collective stakeholder concern alignment. Key findings show that participants’ initial proposals did not maximise perceived utility; negotiated outcomes exhibited weak Pareto dominance over unilateral human proposals in both the utility and cooperativeness spaces; and collaboration perception was grounded on multiple factors. These findings challenge common assumptions in negotiation modelling and provide a representational foundation for future design of collaborative human-robot systems.