Research On Robot Path Planning Based On Ellipsoidal Constraints And Bug Algorithm

Mobile robot should have the ability to change position according to the expectedtarget. Robot path planning is the key of intelligence to decision how to walk in theenvironment. In order to improve instantaneity, adaptability and optimization of robotpath planning in unknown environment, a novel algorithm based on ellipsoidalconstraints and Bug algorithm is presented in this paper, which can overcome theshortage of constraints on obstacles’ shape in the traditional method, and enhance theefficiency of robot obstacle avoidance. The main contents are as follows:Firstly, the research background and status quo of mobile robot are summarized,home and abroad commonly used path planning technologies are reviewed, bycomparing the applied advantages and disadvantages of these methods, the study of thisarticle is put forward.Secondly, the basic theory of convex quadratic constrained quadratic program isexpounded, the working principles and application advantages of ellipsoidal constrainedoptimization model are analyzed. On this basis, the idea of ellipsoidal constrains basedobstacle avoidance path planning is proposed.Thirdly, the path planning model in a two-dimensional static unknownenvironment is established. The path planning in unknown environment is realized inthree steps. First, obtained the obstacles information and determined whether theobstacles and the robot motion would collide. Then, established the ellipsoidalconstrained optimization model, introduced the constraints of obstacles and targetpositions, thus the path planning is equivalent to parameter optimization of ellipsoids.Last, combined this ellipsoidal model with Bug algorithm to calculate the globaloptimal sub-target for solving local oscillation problem, which caused by irregularobstacles.Finally, the ellipsoidal constraints and Bug algorithm based path planning methodis verified. The performance about instantaneity, adaptability and optimization are discussed and compared by a lot of simulations. The results demonstrate theeffectiveness of the proposed algorithm.