Research On Positioning Of UAV Formation Flight

Nowadays,science and technology have developed to a certain extent,the automation of mechanical products has gradually become a requirement of the present era.As a representative product of mechanical automation,UAV is widely used in medical rescue,geological exploration,express transportation and other industries due to their small size,light weight,low cost,reliable performance and flexible application.When a single drone is performing a task,its advantages are limited due to its small load and limited battery capacity.The formation flight of multiple drones can minimize the limitations of the single drone and give full play to the drone's mobility.The precise positioning of the drone is one of the most basic and important prerequisites for exerting its various functions.GPS positioning has become the mainstream synchronous flight positioning method because of its small size and high positioning accuracy,which also causes drones to rely on GPS for synchronous flight positioning.When the GPS signal is artificially shielded or the signal is too weak,the drone's flight function is completely lost.Therefore,after comparing the advantages and disadvantages of signal source positioning,laser positioning,and infrared positioning,this paper proposes the visual positioning method to be used as the positioning method for drone formation flight.Based on the detailed introduction of the flight algorithm,a number of quad-rotor UAV is built to form a set of flight experimental platforms to realize the formation flight of quad-rotor UAV in this paper.For the situation where its positioning is overly dependent on GPS,the camera collects environmental photos in flight,extracts and matches the feature points of the front and back frames.In the experiment,the feasibility of the flight algorithm was verified by Matlab and Simulink simulation,and the feasibility of the flight and SLAM algorithm on the drone were verified by visual simulation of ROS and Gazebo.This paper conducts a single-machine performance test on each drone of the experimental platform to ensure the accuracy of the single-machine visual positioning,and finally performs a synchronous flight test based on visual positioning.The experimental results prove that the vision-based formation flight system designed and implemented in this paper has higher positioning accuracy and better flight formation expectations.This paper has certain application value for UAV formation flight without GPS,and also has certain reference significance for the control of other unmanned mobile robots.