HARDWARE IN THE LOOP SIMULATION FOR A QUADROTOR UNMANNED AERIAL VEHICLE

Abstract

The Unmanned Aerial Vehicles (UAVs) have been well-known as an automatic robot helping people carry out the hard and dangerous works. In the platform configuration, the UAV can be mainly classified into three categories, the fixed-wing types, the helicopter types and the multi-rotor types. The strength of FW types is high speed flight and fuel efficiency. However, it cannot perform to vertical takeoff and landing(VTOL) and hover in the flight itinerary. In the opposite points of view, the helicopter and multi-rotor types can provide the limitation on FW types. By outstanding features of VTOL, a quad-rotor UAV is chosen and used to demonstrate the contribution of this thesis. The main content of this thesis presents the hardware in the loop simulation (HILS) in which the speed communication of the components of HILS is ensure and its can use to simulate the complex flight algorithm and application algorithm such as the vision and the SLAM. Furthermore, a synchronization controller is proposed to improve the performance of the quad-rotor UAV in the tracking trajectory. Firstly, a simulation based on the hardware in the loop is developed for UAV. The HILS includes the pixhawk hardware, gazebo and a middleware software. Therein, the pixhawk is installed a development flight control algorithm by using the PX4 open source. The gazebo is presented the 6DOF, sensor models, 3D visualization and environment. The middle software named the control application software (CAS) is robustly designed to secure communications in the HILS system by using the multithread architecture in Qt Creator, which can to synchronize signals, accelerate time responses and reduce signal losses of the communication between the Gazebo and Pixhawk. Furthermore, the CAS can create line connections to any software and hardware by using TCP, UDP and serial protocol. The CAS provides a GUI that allows users to monitor the status of packet transfer and perform the tasks of a ground control station(GCS) software. Secondly, an accurate tracking controller is proposed for the quad-rotor UAV. By using the modify controller installed on Pixhawk , the angular velocity of each actuator is generated. Hence, a synchronization controller includes the four sliding mode controllers (SMC) to drive velocity actuators and a supervisory controller consists four PID-neural network controllers (PIDNNC) to compensate the synchronization errors between the actuators due to system nonlinearities, uncertainties and external disturbances. The proposed controller can support not only the quad-rotor UAV tracking trajectory but also the 3-RRR parallel manipulator tracking. Lastly, the HILS is extended for the vision applications. For beginning, the camera model is added more to the gazebo to measure the observation environment from the quad-rotor UAV. Next, the raspberry board based Linux OS is put more to HILS setup to perform the vision algorithm. The autonomous landing is applied for the quad-rotor UAV based on the estimate position from the vision algorithm. For advance application of the vision, the graph-SLAM algorithm is developed on the raspberry board. By using the RGB-D camera model on the gazebo like the real Kinect camera, the HILS can demonstrate the SLAM problem which simultaneously localization and mapping for the quad-rotor UAV.