Research And Implementation Of UAV Indoor Autonomous Obstacle Avoidance Based On Vision And Laser

In recent years,with the advancement of science and technology and the continuous reduction of hardware costs,UAV technology has developed by leaps and bounds,and UAV have been widely used in military and civilian applications.However,UAVs still face certain challenges in indoor autonomous flight.It is difficult for UAV to achieve stable positioning and flight in complex environments by relying on a single sensor and being easily constrained by the sensor itself.In this paper,an indoor fourrotor UAV autonomous flight obstacle avoidance system is designed by combining binocular camera and two-dimensional lidar sensor,and the feasibility of the system is tested.The main work of this paper is as follows:Firstly,the visual SLAM(Simultaneous Localization and Mapping)positioning technology is introduced.The AKAZE algorithm is used to extract feature points and the Brute Force(BF)algorithm is used to match.Starting from the stability requirements of UAV positioning,aiming at the problem that BF algorithm is prone to mismatch,RANSAC algorithm is used to remove mismatch to improve visual positioning.Then the Cartographer algorithm of laser SLAM is introduced.In order to solve the problem that the two-dimensional lidar cannot achieve stable positioning in three-dimensional space,the extended Kalman filter is used to fuse the visual pose information and the laser pose information.Then use the laser point cloud to build an octree map,so that the UAV can realize real-time positioning and mapping functions.Secondly,this paper chooses ant colony algorithm for path planning algorithm.Aiming at the problem that the convergence speed of ant colony algorithm is too slow,the artificial potential field method is introduced,and the two algorithms are fused.The pheromone update method and state transition function in ant colony algorithm are improved,and then the simulation and comparison are carried out.It shows that the fusion algorithm has improved the convergence speed and search efficiency.Finally,the whole system framework of the UAV is designed,and the software and hardware system are built.Based on ROS(Robot Operating System)environment,the obstacle avoidance flight of UAV is simulated to verify the effectiveness of the system.Then,the stability of UAV flight is tested by fixed-point flight,and then the obstacle avoidance experiment is carried out to verify the feasibility of the autonomous flight obstacle avoidance system in this paper.