Current predictions indicate that by 2020 the available launchers will not cover the market niche of small satellites. The Aldebaran launcher is intended to cover this market niche. Aldebaran will perform an aerial launch because the initial launcher mass is significantly reduced compared with the ground launch option. Due to the navigation requirements and the precision of the release manoeuvre, Aldebaran will need a hybrid navigation system. Aldebaran will have a strapdown navigation system on board. Gimbaled and floated systems are more precise, but they are heavier and mechanically more complicated than strapdown systems. The sensor output error sources are modelled and the user can specify which parameters are going to be corrected by the navigation algorithm and which not. The random walk is modelled with a dynamic method. The Van Allan variance methods have not been implemented because the validation of this model in some MEMS inertial sensor is still under study. The gravity model used in the navigation algorithm (truncated at the J2 zonal harmonic) is a requirement imposed. However, it has been checked that this requirement is compatible with the required performance. The sensor redundancy has been analysed. We have determined that the Aldebaran launcher will not use ISAs in a non-orthogonal configuration because the sensor output correction algorithm in this case is more complicated than in orthogonal configurations and because the detection of the failed sensor is made by probabilistic analysis. The hybridization corrects the errors of the vertical channel instability and stabilizes it. Therefore, the implementation of an altitude sensor has been discarded because the vertical channel instability problem has been solved. The model has been validated by three sets of tests. The first set validates the attitude determination, the second validates the INS standalone trajectory and finally the hybridization has been validated. The trajectory validation has been executed with real flight data. Finally, the navigation system sensitivity to some parameters has been evaluated with the model. The results show that the accelerometer parameters do not have such a strong influence as the gyroscope parameters. The reason is that, in the navigation algorithm modelled, position and velocity are corrected periodically with GNSS data but the attitude is not corrected. The model implemented in Simulink in this project is the cornerstone of a brand new navigation system model. Such model must be able of evaluating the performance of navigation systems in a wide range of conditions and sensors