Multi-Sensor Data Fusion (MSDF) techniques involving satellite and inertial-based sensors are widely adopted to improve the navigation solution of a number of mission-and safety-critical tasks. Current Navigation and Guidance Systems (NGS) employing MSDF algorithms do not meet the required level of performance in all flight phases of small Remotely Piloted Aircraft Systems (RPAS). Hence in order to satisfy the Required Navigation Performance (RNP), an innovative Square Root-Unscented Kalman Filter (SR-UKF) based NGS is implemented and compared with a conventional UKF design. The presented NGS architectures employ a number of state-of-the-art low-cost sensors including; Global Navigation Satellite Systems (GNSS), Micro-Electro-Mechanical System (MEMS) based Inertial Measurement Unit (IMU) and Vision Based Navigation (VBN) sensors. Additionally, an Aircraft Dynamics Model (ADM), which is essentially a knowledge based module, is employed to compensate for the MEMS-IMU sensor shortcomings in high-dynamics attitude determination tasks. The ADM acts as a virtual sensor and its measurements are processed with non-linear estimation techniques in order to increase the operational validity time. An improvement in the ADM navigation state vector (i.e., position, velocity and attitude) measurements is obtained, thanks to the accurate modeling of aircraft dynamics and advanced processing techniques. A novel SR-UKF based VBN-IMU-GNSS-ADM (SR-U-VIGA) architecture design is implemented and compared with a conventional UKF based design (U-VIGA) in a small RPAS (AEROSONDE) integration scheme exploring a representative cross-section of the operational flight envelope. The comparison of the state vector demonstrates the capability of SR-U-VIGA and U-VIGA systems to fulfill the relevant RNP criteria, including precision approach tasks. Furthermore, the computation time of SR-U-VIGA system is lower when compared to U-VIGA NGS allowing for an enhanced implementation in real-time applications.