Clever Geek Handbook

Investigation of SLAM methods for indoor mobile robot navigation. R2 Robotics Research Experience. (continued)

Introduction

In the last article, we looked at several modern SLAM algorithms for ROS . This article will discuss the application of SLAM in practice. A prototype of a mobile merchandiser robot from R2 Robotics is used as a robot . The robot has a base with two drive wheels located on the same axle in the center, which allows it to make turns on the spot and contributes to high maneuverability. The robot's diameter is ~ 60 cm, and its height is 1.5 meters.





4. Testing

The sensors on the robot are: 2D lidar RPLidar A1 , RGBD camera Intel RealSense D435i and tracking Intel RealSense T265 camera for odometry tracking. The lidar is installed at the bottom of the robot and scans only the frontal sector of 180 degrees, while the camera is set at 1.1 m and tilted downward at an angle of 40 degrees. Considering that the height of the robot is 150 cm, the camera allows you to recognize obstacles at a height inaccessible to the lidar.





Figure 8 - Robot model in RViz
Figure 8 - Robot model in RViz

9. 70 2 . , , .





Figure 9 - Room for mapping tests
9 –

4.1 Rtabmap





– ( 10).





BUT)
)
B)
B)

10 – Rtabmap ) b) RGBD





10 – , , . , . , , , .





, , RGBD , . , - . , , , .





Figure 11 - Comparison of objects displayed on maps: red - racks and high chairs, blue - tables and armchairs, yellow - sofas and wardrobe, green - chairs
11 – : – , – , – , –

11 , , . . , , , 2D . , , ( ). – 7 .





, , . , ( 12).





Figure 12 - Displaying data from an RGBD camera in RViz
12 – RGBD RViz

, , , . , 11 . , , , , .





. , , . , 13 14.





Figure 13 - Avoiding an obstacle on a map built using a depth camera
13 – ,
Figure 14 - Avoiding an obstacle on a map built using a lidar (black - an obstacle on the map, gray - free space on the map, pink - an obstacle on the costmap obstacle layer, blue - inflation layer global costmap, red and blue - inflation layer local costmap)
14 – , ( – , – , – costmap obstacle layer, – inflation layer global costmap, – inflation layer local costmap)

, , , ( 13). , , . , obstacle_layer , ( 14). , . , obstacle_layer . , . , , , «» , , . , , , , .





4.2 Google Cartographer





Google Cartographer . RPLidar A1, Intel RealSense D435i. 





. SLAM, . , , SLAM . , , . . , 2- . , Rtabmap 2 . Rtabmap – SLAM, , Cartographer – , . , 2 , Cartographer . , , , . Cartographer, , , , SLAM.





, , .





Figure 15 - a map built on a lidar using a) Cartographer b) Rtabmap
15 – ) Cartographer b) Rtabmap

15, , Cartographer , , SLAM Rtabmap.





Figure 16 - Visualization of the Rviz map Cartographer, built on the lidar
16 – Rviz Cartographer,

16 . Cartographer. Cartographer , , , , . , - . , Cartographer «» . . , Cartographer – , . , . , ( - ), .





, , . - navigation stack obstacle layer. , . navigation stack, . , , .





Cartographer SLAM SLAM. , Cartographer , , , , . , Cartographer «» , . . , , 5-10 . , ( Cartographer) , , .





, Cartographer , . Rtabmap, Cartographer . Rtabmap , , , , , Rtabmap ( , ).





17 2- , Cartographer. 1, 2, , . , 2 , . , , . Rtabmap , , Cartographer .





Figure 17 - Cartographer 2-room map plotted by lidar
17 – Cartographer 2- ,

SLAM Rtabmap Cartographer , . 2D , . , , obstacle_layer, .





, . -, , 3D . , (, , .) . , , , .





, SLAM. , .





Gmapping – 2D , Cartographer Rtabmap . , Gmapping, 3D .





Rtabmap Cartographer SLAM. Rtabmap , . , , . , – .





Cartographer , ( ) . , SLAM .









  1. Pedrosa, E., L. Reis, C. M. D. Silva and H. S. Ferreira. Autonomous Navigation with Simultaneous Localization and Mapping in/outdoor. 2020.





  2. Gmapping [ ] URL: http://wiki.ros.org/gmapping, – . . . : 14.08.2020 .





  3. Google Cartographer ROS [ ] URL: https://google-cartographer-ros.readthedocs.io/en/latest/#, – . . . : 04.11.2020 .





  4. RTAB-Map, Real-Time Appearance-Based Mapping [ ] URL: http://introlab.github.io/rtabmap/, – . . . : 22.06.2020 .





  5. Adaptive Monte Carlo localization [ ] URL: http://wiki.ros.org/amcl, – . . . : 03.08.2020 .





  6. Building Maps Using Google Cartographer and the OS1 Lidar Sensor [ ] URL: https://ouster.com/blog/building-maps-using-google-cartographer-and-the-os1-lidar-sensor/, – . . . : 25.02.2021 .





  7. Labbé, M, Michaud, F. RTAB‐Map as an open‐source lidar and visual simultaneous localization and mapping library for large‐scale and long‐term online operation. J Field Robotics. 2019; 35: 416– 446.





  8. Silva, B.M.F.D.; Xavier, R.S.; Gonçalves, L.M.G. Mapping and Navigation for Indoor Robots under ROS: An Experimental Analysis. Preprints 2019.





  9. Mathieu Labbé and François Michaud. Online Global Loop Closure Detection for LargeScale Multi-Session Graph-Based SLAM. 2014 IEEE / RSJ International Conference on Intelligent Robots and Systems, pages 2661-2666, 2014.







More articles:


All Articles