The application of the position-position architecture for enabling position synchronization of two robotic agents has been proven effective in the fields of telemanipulation and rendezvous of autonomous vehicles. Nevertheless, the approaches presented to this date with the purpose of rendering the position-position architecture passive under the presence of time-delays and packet-loss are only partially able to fulfil that goal. This owes to the fact that they mostly focus on passivating the system, at the cost of transparency. Such an issue becomes even more critical in the presence of position drift caused by most passivation methods. This letter presents a novel control approach that enhances the position synchronization of agents suffering from delayed coupling, by introducing a local proxy reference to one of the agents and only closing the feedback loop when it can preserve stability. The concept is free of position drift and promises less conservatism, without having any prior information about system parameters or prior assumptions regarding time-delay. It has been experimentally validated for time-varying round-trip delays of up to 2 s.