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. Such integrated Navigation and Guidance Systems (NGS) currently do not meet the required level of performance in all flight phases of small Remotely Piloted Aircraft Systems (RPAS). In this paper an innovative Square Root-Unscented Kalman Filter (SR-UKF) based NGS is presented and compared with a conventional UKF governed design. The presented system architectures adopt state-of-the-art information fusion approach based on a number of 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 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. An innovative SR-UKF based VBN-IMU-GNSS-ADM (SR-U-VIGA) architecture design was implemented and compared with a typical UKF design (U-VIGA) in a small RPAS (AEROSONDE) integration arrangement exploring a representative cross-section of the operational flight envelope. The comparison of position and attitude data shows that the SR-U-VIGA and U-VIGA NGS fulfill the relevant RNP criteria, including precision approach tasks.