Research On Attitude Measurement And ROS Software Design For Rotor UAVs

Rotor UAV has the characteristics of compact structure,strong maneuverability and flexible operation.It can take off and landing vertically and hover in the air with a variety of airborne equipment.It is widely used in military,agriculture,film and television fields.However,in order to achieve the smooth flight of the Rotor UAV,complex sensors,fast controllers and actuators and complex navigation control algorithms are needed as support.And pose information is like the "eyes" of UAV,which is the basis for UAV to achieve autonomous flight.As a multi-sensor data fusion system,INS/GPS integrated navigation system can obtain high-precision navigation information of UAV,and attitude information is the premise of navigation.Relying on two important subjects of the project team,this paper extensively studies attitude fusion of Rotor UAV based on multi-sensor data fusion,and compiles corresponding software using ROS system in UAV.The main work is as follows:(1)Aiming at the task requirement and working environment of the system,the hardware experiment platform of the Rotor UAV is built.The performance indicators of related hardware modules are analyzed.Pixhawk is used as an open source flight control programmer to write the corresponding program for secondary development,and QGC is used to adjust the parameters such as control.The control structure block diagram and software workflow diagram of the system are designed.(2)As the theoretical basis of this paper,the complementary filter algorithm is studied in depth.The basic principle of complementary filter is analyzed from the point of view of time domain and frequency domain.Through the relationship between complementary filter coefficient and filter turning frequency,an improved method of complementary filter is proposed to cope with the change of system noise characteristics.Then the method of attitude fusion using gradient descent method is studied,and the shortcomings of the algorithm under the interference of convergence speed and non-gravity acceleration are analyzed.On this basis,the attitude-based data fusion scheme and the improvement bill are determined.(3)Studying the basic principle of quaternion attitude calculation,using quaternion to update attitude,designing a quaternion attitude calculation method based on complementary filtering;using the relationship between acceleration and attitude angle,using gradient descent method to design attitude fusion algorithm.Considering the non-gravity acceleration of UAV and the slow convergence speed,the gradient descent algorithm is improved.A Nesterov accelerated gradient attitude fusion algorithm is proposed.Finally,the feasibility of the algorithm is proved by comparison experiments.(4)Install and use ROS on the airborne computer.MAVLink protocol is used to solve the communication problem between ROS system and Pixhawk,and the connection structure diagram between nodes and topics of ROS control UAV is analyzed.The SITL simulation of UAV in Gazebo environment is realized by compiling ROS node program,and the experimental results are observed by QGC.Finally,the feasibility of ROS software for UAV control is verified by experiments on the test platform of Rotor UAV.