Reaching accurate state estimation is necessary for the unmanned aerial vehicle (UAV) to conduct an efficient flight characterized by exceptional stability. Achieving reliable state estimation is regarded challenging because of the errors related to the measurements of the onboard MEMS (Micro Electro-Mechanical Systems) which constitute IMU (Inertial Measurement Unit). The UAVs massive rotors vibration causes enormous drift, biases and unexpected noise sequences which makes these measurements inaccurate. Applying traditional estimators will compensate these errors but less computationally intensive techniques should be created. In this paper, a deep learning (DL) framework based on deep neural networks (DNNs) is proposed to estimate UAV attitude estimation. Batch Normalization and dropout techniques are implemented for training to prevent overfitting and decrease the computational overhead of nets. The suggested DL technique proves the ability to use trained DNN. Furthermore, Euler kinematical equations model were developed, and synthetic data samples were collected to train the DNN. Finally, a comparison study between the proposed approach and the kinematical model is conducted. The results indicate that the suggested method can successfully replace difficult approaches for calculating and measuring attitude. This approach has the potential to improve the performance of UAVs and reduce the costs of state estimation. (This study was produced from the master's thesis of the first-ranked author at Marmara University. ORCID NO: 0000-0002-5381-9736)
Anahtar Kelimeler: Attitude Estimation, Deep Learning (DL), Batch Normalization, Dropout Technique, Quadrotor.