The Key Technology Of Field Autonomous Navigation Robot Multi-sensor Information Fusion

In recent years,mobile robots have developed rapidly and replaced professionals to carry out relatively dangerous tasks in the wild environment,such as mountain cave environmental detection and post-earthquake debris search and rescue,which have become the mainstream.However,the field environment is relatively complex and requires high requirements for autonomous navigation.In order to make mobile robots adapt to a variety of complex environments in the field,the key technology of multisensor information fusion for autonomous navigation robots in the field is studied in this paper.In order to perform tasks in the complex environment in the field,mobile robots should first be able to perceive the complex environment in the field.Binocular camera and lidar have high detection accuracy and are suitable for mobile robot to sense the external environment.Therefore,this paper USES two kinds of sensors: binocular camera and lidar.However,in the process of using binocular cameras,the light requirements are high while the point cloud data detected by lidar is relatively sparse.In order to reduce the detection error and complement the advantages of the two sensors,this paper proposes a method of data fusion based on point cloud.First of all,by binocular camera to obtain three-dimensional point cloud into pseudo two-dimensional lidar point cloud,and then the laser radar for 2 d point clouds and pseudo lidar point cloud using filter outliers,and then use the CIF descriptor method to make two point cloud registration,then make two point cloud by using the method of weighted average fusion as a point cloud,finally using MATLAB simulation and experiment of this method by the simulation results show that the fusion of the point cloud noise was significantly reduced.The awareness of the external environment is only the first step to complete autonomous navigation,and high-precision positioning technology is also needed.In order to enable the mobile robot to carry out high-precision navigation,it is necessary to obtain the pose information of the mobile robot.In this paper,IMU and wheeled odometer are used to obtain the pose information of mobile robots,and EKF algorithm is used to fuse the information,and point cloud information is used to correct the accumulated error.The acquisition of pose information and point cloud information lays a foundation for the path planning of mobile robots.In this paper,the advantages and disadvantages of the current commonly used path planning methods and drawing methods are analyzed in detail,and MATLAB is used to simulate them.Then according to the characteristics of the field environment,the raster method is selected to construct the map.Dijkstra algorithm is used for global path planning.DWA algorithm is used for local path planning selection.After the theoretical analysis is completed,the mobile robot platform is built to program and test the method proposed in this paper.The experiment proves that the mobile robot can achieve high-precision autonomous navigation and positioning in a relatively complex environment through a series of algorithms described in this paper.