Research On Indoor Navigation And Control Technology Of Quadrotor UAV

Quadcopter UAV is currently widely used in aircraft,in this intelligent era,quadcopter UAV research focus on autonomous navigation,for indoor environment global positioning system failure,resulting in quadruple UAV indoor positioning inaccurate problem,this paper designs a quadcopter UAV for exploring indoor unknown environment,build a set of autonomous navigation system,instead of manual to deal with the complex operation of indoor environment.The main research contents are as follows:(1)Develop the system scheme of indoor quadcopter UAV,design from the hardware platform and software system of indoor quadcopter UAV,first complete the hardware selection calculation and hardware platform construction,and then conduct the principle analysis and camera calibration experiment of the camera model for the visual sensor in the hardware,the software first studies the control technology of the quadcopter UAV,analyzes the flight principle of the quadcopter UAV,and uses MATLAB to write the PID closed-loop controller of the quadcopter UAV.After parameter adjustment,the control simulation experiment by the robot operating system ROS verifies the rationality.(2)Aiming at the problem that the mainstream visual SLAM algorithm ORB-SLAM2 in indoor navigation has uneven features leading to the loss of severe motion tracking,the ORB-SLAM2 algorithm is improved,a front-end feature extractor based on object detection is designed,and the improved algorithm can improve the accuracy and robustness by conducting simulation experiments of the dataset in the ROS system.(3)Aiming at the problem that the mainstream 3D reconstruction algorithm Elastic Fusion in indoor navigation has the problem that the 3D map details are not fine,the implicit mapping algorithm Instant-ngp is selected to optimize the real-time reconstruction of the Instant-ngp algorithm,and the dataset simulation experiment is carried out through the ROS system,and the 3D map obtained by the Instant-ngp algorithm is finer than the mainstream 3D reconstruction algorithm Elastic Fusion.(4)For the path planning in indoor navigation,the EGO-Planner algorithm with high planning efficiency is adopted,and the feasibility of the algorithm is verified by the path planning simulation experiment in the ROS system and the indoor path planning experiment in the real environment.(5)The navigation simulation experiment in the ROS system and the indoor navigation flight experiment in the real environment were completed,which verified the feasibility of the simulation platform and physical navigation.