In the coal mine roadway 、 Subway tunnels and other dangers 、 Under the claustrophobic underground scene , Use the mobile robot to complete the detection 、 Mining and search and rescue missions are safe and efficient . The underground robot can accomplish the task autonomously and intelligently , Accurate positioning and map construction are the premise and key .
chart 1- Robots in underground tunnels
When robots run autonomously in the underground environment, they often have no prior map information , And it can't be used GPS Positioning , Robots are required to create maps in unknown environments , At the same time, the map is used for autonomous positioning and navigation , namely SLAM technology .
Because the scenes in the underground environment are very similar in geometric characteristics , And the LIDAR point cloud is rarely distributed in the distance , Based on lidar SLAM The method is not effective ; Researchers from China University of mining and technology have adopted a multi-sensor fusion method , The framework based on graph optimization will UWB and IMU The position constraints provided by the fusion positioning system are added to the pose graph optimization constraints , Provide reliable initial estimation for lidar scanning matching , Multiple sensors cooperate to estimate the state of mobile robot .
Firstly, a fusion algorithm based on extended Kalman filter is proposed UWB Ranging information and IMU Algorithm of acceleration information , By augmenting the state vector , The acceleration and the deviation of acceleration are also estimated , It can improve accuracy and reduce delay , It can provide reliable location estimation for mobile robot in underground environment .
In order to verify the effectiveness and positioning accuracy of the algorithm , And the practicability of the long and narrow tunnel environment , The researchers designed an indoor validation experiment . Experimental use Turtlebot2 Mobile robot as a robot platform , And fixed on the platform IMU and UWB Mobile nodes .UWB Use 4 Anchor nodes to build a positioning system .
chart 2- Mobile robot platform
Layout around the site 8 individual NOKOV Measure Mars2H Motion capture lens , utilize NOKOV The measurement motion capture system tracks the reflective identification points pasted on the mobile robot , To get the real trajectory of the robot .
chart 3- Experimental scenario
by force of contrast NOKOV Measure the reference track collected by the motion capture system （ The real track ） and EKF The estimated trajectory output by the algorithm can be seen , The estimated value is basically consistent with the real value .
chart 4- Real trajectory and estimated trajectory
After verifying the performance of the above positioning scheme , Researchers have developed a laser in the environment of long and narrow underground tunnels / Ultra wideband convergence SLAM Algorithm , And the experiment is carried out in the actual underground tunnel , It is proved that the method is closer to the actual trajectory and there is no cumulative error .
reference ：[1] Zhao Yu . Research on location and mapping method of mobile robot for underground long and narrow tunnel [D]. China University of mining ,2021.
