Study Of Excavating Robot's Autonomous Navigation And Obstacle Avoidance Based On VSLAM

With the rapid development of computer technology and sensor technology,it provides technical support for the progress of excavation robots.At the same time,with the rising labor costs,there is a growing demand for excavators that can work independently.The autonomous obstacle avoidance navigation of excavating robot is a very important part of realizing intelligence and cluster control.This paper designs an autonomous navigation system for excavating robots,which can achieve simultaneous mapping and positioning.Operators can select target points on the upper monitor,and the excavating robots can move to the target position independently.In addition,the visual odometer based on multi-sensor fusion can ensure the normal positioning of the system when part of the sensor information is lost.It improves the accuracy and robustness of the system.In view of the large-scale working environment of excavating robot,octree map can save a lot of storage space,and octree maps with different resolutions can be set according to different working scenes.The main contents are as follows:In chapter 1,the significance of autonomous navigation system of excavating robot is expounded,the related research status is introduced,and the research content of this paper is put forward according to the research goal of this paper.In chapter 2,Designing a navigation obstacle avoidance system for the needs of excavating robots.Models used in SLAM are described mathematically.It provides a theoretical basis for the subsequent algorithm research and implementation.In chapter 3,The VSLAM algorithm based on RGB-D camera is studied.Then the visual odometer,back-end optimization,loop detection and map building are introduced in detail.The construction of point cloud map,octree map and grid map is realized.In chapter 4,The models of Kalman filtering algorithm and extended Kalman filtering algorithm are introduced and deduced.Finally,the framework of visual odometer based on extended Kalman filter is introduced.The odometer based on multi-sensor fusion improves the robustness of the system and can obtain better pose estimation at both high and low speeds.In chapter 5,The navigation system of excavation robot based on ROS is built.The principle and process of global path planning algorithm A*and local path planning algorithm DWA are studied.The Gazebo model of the excavating robot is established and the corresponding sensors are added.At the same time,the relevant environment is established as the simulation environment.In chapter 6,The positioning performance of the odometer based on multi-sensor fusion is verified by linear and circular experiments.The feasibility of autonomous navigation obstacle avoidance for excavating robots is verified by experiments and simulations of autonomous navigation obstacle avoidance.And verify the feasibility of multi-robot autonomous navigation obstacle avoidance in the simulation environment.In chapter 7,This paper summarizes the main contents of the paper,and finally puts forward the ideas and ideas for further development on the basis of this paper.