In recent years there has been an increasing interest in mobile multi-agent systems, with the problem of how to drive a formation of N agents from one configuration to another being a key issue in many applications, particularly when each agent has limited (communication and sensing) capabilities. One such example is the navigation of glider formations, where N AUVs have to reach a target position within a given time, with sensing and communication being limited to discrete time instants. With this example as motivation, we begin by studying the single agent version of the problem, where reachability concepts are used to deal with the issue of position uncertainty, allowing us to develop two different control strategies. These strategies are then used to tackle the two-agent problem, which we generalize to obtain a partially decentralized approach to the multiagent problem. Sufficient conditions for the successful application of these strategies, as well as the corresponding bounds on the uncertainty are also derived. The problem of estimating the disturbance set is then studied, with the objective of providing the network with increased robustness to scenarios in which there is few information about the disturbance set. Finally, using normally distributed disturbances, some relevant scenarios of both the single and multi-agent problems are simulated, which confirm the results.