Research On Quadcopter Three-Dimensional Obstacle Avoidance Algorithm Based On AirSim Simulation

With the advancement of science and technology and the rapid development of economy,the figure of quadrotor gradually appears in various fields of production and life.The three-dimensional motion characteristics of the quadrotor and the complexity of the actual three-dimensional scene also greatly expand the development direction of the quadrotor in the field of automation.Based on this,the paper will carry out autonomous obstacle avoidance related research on the three-dimensional environment of quadrotor quadrotors in unknown maps and suburban residents’ living area types without GPS signals,Obstacle avoidance system operated using AirSim by the obstacle avoidance algorithm is constructed.The improvement and optimization of the local optimization problem and the algorithm instability problem caused by special scenes in the algorithm finally achieve the safe and stable quadrotor reaching the specified target point.The paper completes the construction of a quadrotor flight simulation platform,analyzes the current simulation principles and architecture characteristics of each platform,and studies and analyzes the main simulation modules in airsim according to the constraints required by the subject.The solution combined with ROS system solves the problem of large delay in information acquisition and visualization in AirSim,and then realizes the fixed-point hovering and stable flight of a specific trajectory of the quadrotor quadrotor,which satisfies the unmanned simulation in this study System requirements.Based on the simulation platform,this paper designs a set of quadrotor obstacle avoidance control system that meets the research needs.Using binocular stereo vision as the perception,the quadrotor real-time positioning and specific three-dimensional map obstacle avoidance system are constructed.This system mainly uses the visual SLAM method to estimate and position the quadrotor.It uses a three-dimensional map model of the depth distance information to build,and finally completes a set of obstacle avoidance schemes suitable for the research of this subject.Obstacle algorithm research and experimental verification laid the foundation.On this basis,the paper studies the obstacle avoidance algorithm of the quadrotor in the three-dimensional unknown suburban residents’ living environment,and analyzes the current obstacle avoidance algorithms according to the factors such as no GPS,outdoor lighting,obstacle types,geometric characteristics,etc.The scope of application is studied,and the deployment of the 3DVFH algorithm in the local obstacle avoidance on the quadrotor is studied.The problem in which the algorithm falls into the local optimal solution is mainly studied,and the problem is solved from three aspects.A histogram generation algorithm with memory features,designed a strategy to weaken the "reciprocating" flying degree of the quadrotor,using the dynamic coefficient method to improve the cost function,thereby solving the problem of drone falling into the local minimum cycle.Combining these three aspects,a comparative experiment on the improved 3DVFH obstacle avoidance algorithm with memory features is carried out.Finally,the improved algorithm is carried out by using the constructed quadrotor obstacle avoidance system scheme on the completed drone simulation platform.This paper confirmed the stability and superiority of algorithm.