Research On 3D Terrain Environment Perception Modeling Of Mobile Robot Based On Local Composition

With the development of the economy and the advancement of technology,society has put forward new requirements for the development of mobile robots,namely,intelligence.The premise of intelligent realization of mobile robots is their autonomy,and the key to autonomy is the ability to perceive the environment.At present,the development of indoor mobile robot environment sensing technology has become increasingly mature,while outdoor mobile robots are affected by many factors,and the development speed is slightly slow.The range of activities of outdoor mobile robots is increased,and the amount of environmental information is increased.If the information storage is too complete,the storage cost will increase,and if it is too rough,it will be difficult to use it for application tasks such as robot path planning and obstacle avoidance;Under environmental conditions,it is impossible to set reference objects such as beacons and landmarks.The cumulative error caused by the long-term operation of the pose sensor is difficult to eliminate,which makes it difficult to meet the absolute positioning accuracy.In view of the problems of outdoor mobile robots in terms of environment perception,this paper mainly completed the following research work:(1)Design 3D terrain environment storage scheme,it is difficult to accurately describe 3D environment information for 2D grid map,and 3D grid map data processing efficiency is low.This paper improves the 2D grid map as follows: Consider terrain elevation information Replaces the probability information of whether the point stored in the two-dimensional grid is occupied.It can meet the requirements of data storage efficiency and visually describe the three-dimensional terrain environment.(2)A three-dimensional terrain environment mapping algorithm for mobile robot based on local composition is proposed.By establishing a three-dimensional model of the terrain environment in a limited area around the robot,as a description of the environment details when planning the global path of the robot,the storage cost can be effectively saved.The "robot-centric" mapping method weakens the influence of the robot's pose drift on the composition and reduces the dependence on the absolute positioning accuracy of the robot.(3)Build the overall framework of 3D terrain environment awareness system,use ROS system as software platform,program the 3D terrain environment mapping algorithm,and verify the effectiveness and practicability of the algorithm in simulation and real environment respectively.