GPS navigation system is widely used for terrestrial vehicle navigation system. Nevertheless satellites visibility can be interrupted by city obstacles such as trees, tunnels or urban canyons. Nevertheless, the trajectory can be partially reconstructed by means of inertial measurements units but, while the integrity of navigation data signal is provided in Europe using the EGNOS system by means of the Horizontal Protection Level calculation, when GPS data are missing there is an impossibility to associate an HTL (Horizontal Trust Level) to position measurement. This paper deals with the possibility to use the covariance matrix propagated in a Kalman filter to obtain an evaluation of the HTL. Some sensors classes, with different accuracies and in different configurations have been analyzed and numerical results are described. Moreover a real city path has been performed, acquiring GPS and Inertial Measurement Unit data, and trajectory reconstruction data are shown and commented, paying particular attention to differences between simulations and “real-world” cases.