The IAR Twin Otter Atmospheric Research Aircraft has a continuing requirement for more accurate, inertially-based navigation data for both track recovery and the calculation of wind gust components. This navigational accuracy is necessary, not just during post-flight analysis, but also for realtime, in-flight guidance and wind computation. Previous developmenlal work at the Flight Research Laboratory on advanced navigation systems has demonstrated the benefits of a Kalman filter integrated navigation approach in order to satisfy the most stringent navigational requirements. A significant upgrade to the navigation sensor suite onboard the Twin Otter in the last two years has resulted in the potential, via Kalman filtering, for generating very high quality inertial velocily and positional information in real time, together with improved airborne wind components. The Kalman filter integrated navigation design described in this report is based on the optimal blending of data from an LTN-90-100 strapdown Inertial Reference System (IRS), a Decca Type 72 Doppier velocity sensing (DVS) system and an ARNAV R-40 airborne Loran-C receiver - sensors that are available on the Twin Otter at the present time. In the Twin Otters real-lime computing/data acquisition system, all three of these navigation sensors are interfaced to the onboard LSI-11/73 microcomputer system, and a complete set of navigation parameters is being recorded. In particular, all of the raw inertial data parameters from the LTN-90-100 IRS, required for proper design of an IRS-based Kalman filter, are available with sufficient resolution and at a suitable digital sampling rate. A driving force behind the decision to pursue this integrated navigation approach on the Twin Otter has been the observation that significant velocity errors (and, eventually, position errors) can occur in the LTN-90-100 IRS over the course of a flight, and the observed error levels can seriously degrade the accuracy of the wind calculations. On the other hand, the airborne Loran-C positional data has been demonstrated to be consistently more accurate than the IRS position information, in the long term. An integrated navigation system approach, using the principtes of Kalman filtering, is shown to have the ability to use Loran-C data (and, to a lesser extent, Doppler velocity data) to accurately track the dominant IRS errors (position, velocity and attitude components), and provide IRS error corrections at a rate appropriate for Twin Otter requirements.