In multi-agent systems, most commonly, consensus consists in all agents converging to a common equilibrium point. For nonholonomic vehicles, which move on a Cartesian plane and occupy a physical space, it is more appropriate to speak of consensus-based formation, which implies that the vehicles are required to converge to a relative position in a given formation while, distributively, agreeing on the center of such formation and on their orientation. In this paper we solve such a problem via a smooth, time-varying, output-feedback control; more precisely, without velocity measurements. In addition, and this is our main contribution, the controller is designed to satisfy pre-imposed bounds to avoid input saturation. To the best of the authors' knowledge this is the first controller of its kind in the context described.