An Integrated Navigation System Design For Quadrotors

Abstract

In this work, a particle filter based integrated navigation system is designed and tested with real quadrotor systems. Conventionally, Kalman filter based algorithms is used for navigation systems which is highly non-linear. Kalman filter has significant assumptions such as linearity of system and measurement models and Gaussian distributions. This situation causes that Kalman filter does not work effective with navigation systems. Nowadays, particle filter based algorithms take over with respect to other estimation algorithms because of progression of computer's processors capability. Within the scope of this study, a particle filter based integrated navigation system is designed. This designed navigation system is tested on real quadrotor system. As a result of the real flight tests, the error margin is reduced a quarter according to the most trusted navigation aid.