Fusion of inertial measurement system and visual navigation for unmanned aerial vehicles using Hardware in the Loop simulation

The main purpose of this paper is presentation a novel approach of inertial-vision based navigation to increase accuracy and safety factor simultaneously. In order to estimate navigation parameters, consist of velocity, position and orientation (attitude) the Extended Kalman Filter (EKF) is used. In proposed method, the Inertial Measurment Unit (includes of 3-axis accelerometer and gyrospope) and vision data considered as navigation process model and measurement model respectively. By utilizing the Scale Invariant Feature Transform (SIFT) method, means transform the image data in to scale invariant coordinates relative to local feature, the 2D position of UAV is obtained. The real time Hardware in the Loop (HIL) simulation technique is used To evaluate overall system performance and efficiency such as serial interface time delay, appropriate operating frequency selection and micro-controller performance at implemented image processing and data fusion algorithm. The real time HIL simulation results clearly show despite of low frequency in vison based navigation, proposed approach has acceptable accuracy for estimating uav navigation paraqmeters. Maintaining accuracy and cost, the proposed approach can be a suitable alternative method to drones that use GPS-based navigation.