Localization systems based on ultrawide bandwidth (UWB) technology have been recently considered for indoor environments, due to the property of UWB signals to resolve multipath and penetrate obstacles. However, line-of-sight (LoS) blockage and excess propagation delay affect ranging measurements thus drastically reducing the localization accuracy. In this paper, we first characterize and derive models for the range estimation error and the excess delay based on measured data from real ranging devices. These models are used in various multilateration algorithms to determine the position of the target. Using measurements in a real indoor scenario, we investigate how the localization accuracy is affected by the number of beacons and by the availability of priori information about the environment and network geometry. We also examine the case where multiple targets cooperate by measuring ranges not only from the beacons but also from each other. An iterative multilateration algorithm that incorporates information gathered through cooperation is then proposed with the purpose of improving the localization accuracy. Using numerical results, we demonstrate the impact of cooperation on the localization accuracy.