Path Planning And Navigation Based On Improved Genetic Algorithm Under Mutiple Constraints

With the development of science and technology,mobile robots have been active in various application scenarios,such as home environment,restaurant and so on.In practical applications,robots often need multiple constraints when they walk.The improvement of one constraint will lead to the loss of the other,so it is difficult to make a choice among multiple constraints.Therefore,The application of path planning and navigation based on improved genetic algorithm in mobile robot is studied in this thesis.Firstly,through in-depth understanding and analysis of the mainstream navigation sensors and the commonly used mobile robot navigation technology,visual sensors and genetic algorithm are used on the Robot Operating System(ROS)to complete the overall scheme design of mobile robot navigation system based on ROS under multiple constraints.Secondly,aiming at the problem of large computation and inconsistent map construction when Simultaneous Localization and Mapping(SLAM)is easily generated by using visual sensors,a closed-loop detection algorithm based on improved visual dictionary tree is proposed.First of all,the color feature,point feature and line feature of the image are extracted.Then,the color feature of the image is used to coarse filter the image,and the visual dictionary tree with the fusion of point feature and line feature of image is used to fine filter the image,and the line segment detector(LSD)feature extraction algorithm and image similarity calculation formula are improved to get the candidate closed loop frame.Finally,add closed-loop constraints to the filtered image.The experiment shows that the improved closed-loop detection algorithm proposed in this thesis can estimate the more accurate posture of the robot and meet the real-time requirements of the system.Thirdly,aiming at the disadvantages of low search efficiency and slow convergence speed of in the mobile robot path planning under multiple constraintconditions,a path planning algorithm of mobile robot based on improved genetic algorithm under multiple constraint conditions is proposed,which fully considers the influence of path length,smoothness and safety.by analyzing the shortcomings of genetic algorithm in initializing population under multiple constraints,the algorithm produces points around obstacles to generate the initial path by surrounding point set(SPS)algorithm,so as to improve the ability of the algorithm to generate initial population quickly.Then,smoothing operator and deleting operator are added to delete unnecessary points and make the path smoother.Finally,the niche method is combined to maintain the diversity of population and avoid the phenomenon of prematurity.The experimental results show that the improved algorithm has some advantages in path length,path smoothness and path safety,and the convergence speed of the algorithm is also slightly improved.Finally,the improved closed-loop detection algorithm and the improved genetic algorithm under multiple constraint conditions are integrated on the ROS platform,and apply to the autonomous navigation of mobile robot.Through the related experiments,it is proved that the path planning and navigation based on improved genetic algorithm is feasible and stable in the field of mobile robot.