State estimation of longitudinal state vectors of unmanned aerial vehicle using kalman filter and luenberger observer

In today's world, interest in unmanned aerial vehicles has ben constantly increasing. Unmanned aerial vehicles have become an important research subject in both military and civilian areas due to their flexible mobility independent of the environment, vertical landing and vertical take-off capabilities without the need for runways, and providing high accuracy, time and cost savings to the user according to their usage areas. Due to the lack of a driver, these dynamic systems need to have an effective control mechanism. To provide effective control, it is necessary to accurately estimate the state vectors that manage system dynamics. Thus, the future control inputs of the aircraft are adjusted in the most accurate way so that the system has real-time measurement data. However, it is not that easy to estimate the state vectors of a dynamic system. Noisy measurements from the system, uncertainties in the system model, and complexity of the system make it difficult to estimate state vectors. A variety of methods are available in the literature for state estimation. In this thesis, linear version of Kalman Filter and Luenberger Observer were selected for estimating angle of attack, pitching rate, air speed, angle of pitch state vectors of unmanned aerial vehicle expressed by linearized longitudinal state space representation. Firstly, the longitudinal state space representation of the unmanned aerial vehicle is modeled in MATLAB environment and then the filter and observer are designed in MATLAB and Simulink environment. For the designed filter and observer, both input signals with fixed values and different frequency sine signals are given to the system. Later, simulations were made for both Kalman Filter and Luenberger Observer in a noisy and noiseless environment. Finally, the performance of the Kalman Filter and Luenberger Observer to estimate the state vectors of the modeled aircraft were compared. According to the simulation results, Kalman Filter shows higher performance than Luenberger Observer in stochastic environment to estimate the state vectors of aircraft, but Luenberger Observer is much more successful than Kalman Filter in the absence of irregularity in the environment.