Research On Velocity Adaptive Motion Planning Of Quadrotor

With the advantages of high flexibility,high mobility,and simple structure,quadrotors are widely used in military and civilian tasks such as reconnaissance and search,mine exploration,logistics and distribution.Driven by the era of automation,intelligence,and digitalization,various industries have put higher requirements for the autonomy and safety reliability of quadrotors.How to build a motion planning system to generate collision-free,smooth,and executable high-quality trajectories has become a challenging problem that needs to be urgently researched and solved.At present,although a large number of researches on quadrotor motion planning have emerged,most of them do not consider the impact of environmental safety margin on flight velocity,which makes it difficult to achieve a balance between safety and efficiency.In view of the above,this thesis,oriented to the adaptive adjustment of velocity with the environment,proposes three motion planning algorithms based on quadratic programming,unconstrained optimization,and Euclidean distance field free to achieve safe and reliable autonomous flight of quadrotors.The specific work of this thesis is as follows:(1)Quadratic programming based motion planning with velocity adaptation.Firstly,an improved A*algorithm with adaptive node expansion is proposed for the initial path search,based on which a safe flight corridor without obstacles is generated.Secondly,piecewise Bezier curves are used to describe the trajectory,and the trajectory optimization is formulated as a typical quadratic programming problem.To achieve velocity adaptation,we estimate the environmental safety margin near the trajectory based on the safe flight corridor,and take the safety margin as a factor affecting the velocity cost.Thirdly,in order to make the velocity adaptation strategy applicable to scenarios with sparse obstacles,a posteriori safety detection method is proposed.Finally,the effectiveness of the proposed algorithm is verified by simulation experiments and physical experiments,i.e.,the velocity can be adaptively adjusted with the environment.(2)Unconstrained optimization based motion planning with velocity adaptation.B-spline curve is used for mathematical description of the trajectory due to its excellent nature such as local modifiability and convex hull property.Based on the nature of B-spline and the optimization objective of the trajectory,we construct an unconstrained optimization problem framework.In this framework,a velocity cost function that can be adaptively adjusted with the environment is designed using the obstacle distance and distance gradient provided by the Euclidean distance field.After solving the unconstrained optimization problem,the trajectory is checked for safety and dynamic feasibility,and a new time allocation method is proposed from the perspective of velocity adaptation.Finally,experiments on autonomous obstacle avoidance of quadrotors with limited sensing range are conducted in the simulation platform to verify the effectiveness of the proposed algorithm.(3)Euclidean distance field free based motion planning with velocity adaptation.In order to reduce the computational cost of motion planning,the gap between the trajectory inside the obstacle and the collision-free path is used instead of the Euclidean distance field to obtain the obstacle distance information and distance gradient information.On this basis,we construct the collision cost function and use the collision cost as the influence factor of the velocity cost,so as to propose a new adaptive adjustment scheme of velocity with environment.The effectiveness of the proposed algorithm is verified in simulation experiments and real quadrotor flight experiments.