Accurate estimation of the pose and velocity of a drone is often critical to the success of its mission. While INS/GNSS methods provide satisfactory results when a clear reception of GNSS signals is available, they quickly fail in case of loss of the satellite positioning signals. On the other hand, Vehicle dynamic model (VDM) based navigation provides better resilience to GNSS outages, by fusing drone aerodynamics with the available onboard sensors, however, this approach has so far been developed only using single IMU measurements. Although robustness to GNSS outages has been well demonstrated in previous works, the notion of handling IMU failures has not yet been addressed in the context of model-based drone navigation. As a first step in this direction, we investigate the feasibility of fusing multiple IMUs within a model-based system, without entering the domain of fault identification and isolation. Our results demonstrate that the use of redundant IMUs is indeed realizable in the context of VDM-based navigation. Besides, we observe that the addition of multiple IMUs does not further alleviate the already improved navigation accuracy by the VDM. However, the proposed framework shows a substantial potential to enhance the reliability of the system in case of IMU failure. Finally, we measure the increase of the computational burden caused by the algorithm, and show that its real-time implementation is still feasible on modern onboard computers.