This paper presents a distributed maneuver planner for connected automated vehicles (CAVs) built upon the idea of CAVs being able to explicitly share their intentions by creating virtual vehicles (Us) on the targeted traffic stream. Specifically, this work focuses on the decision-making process concerning the merging maneuver at single-lane roundabouts, for which a traffic model is exploited to assess the suitability of potential merging gaps. Au optimization-based trajectory planner is then used to generate collision-free, smooth trajectories to successfully perform the planned maneuver. The algorithm is tested in simulation and its performance is compared with a gap-acceptance-based driving policy. The results show the potential positive impact that the proposed VV-based interaction mechanism along with the traffic-model based cooperative maneuver planner could have on traffic coordination.