研究目的
Comparing and evaluating three 2D-SLAM algorithms based on laser radar in the robot operating system (ROS), namely Gmapping, Hector-SLAM and Cartographer, to determine their strengths and weaknesses in different environments.
研究成果
Gmapping has the highest map accuracy in simple small scene environment, Hector-SLAM is more suitable for corridor type environment, and Cartographer has advantages in complex environment. Due to the limitations of the actual environment, the experimental environments selected in this paper are relatively small.
研究不足
The experimental environments selected in this paper are relatively small. Therefore, in the following work, on the basis of the existing experimental platforms, large experimental environments with structural characteristics will be selected for SLAM mapping experiments, so as to facilitate the analysis of the advantages of the Cartographer algorithm in back-end optimization.
1：Experimental Design and Method Selection:
Built a mobile robot experimental platform based on ROS in the real environment. Compared three 2D-SLAM algorithms: Gmapping, Hector-SLAM, and Cartographer.
2：Sample Selection and Data Sources:
Experiments were carried out in a simple corridor and a laboratory with many obstacles. Ten points in the real environment were selected to measure the distance on maps and the real distance obtained by laser range finder.
3：List of Experimental Equipment and Materials:
RPLIDAR A2 laser radar, NVIDIA Jetson TX2, STM32F1 driver board, GY-85 nine-axis gyroscope, large load encoder motor, and programming operation platform.
4：Experimental Procedures and Operational Workflow:
The robot's linear velocity was set as
5：3m/s and the angular velocity was 5rad/s. The algorithms took the default parameters. Rviz was used for map visualization. Data Analysis Methods:
The distance on the map was measured through the measure tool on the RVIZ, and then compared with the actual distance measured by the laser rangefinder for error analysis.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
