Comparison of Low-Cost Odometry for Mobile Robot in Various Indoor Environments

---

Overview

• Paper

• Presentation Video

• Presentation Slide

Our Dataset

• .rosbag file download link

Scenario Files	Features	Object Type	Driving Path Pattern
s1.bag	Many	Static	Straight
s2.bag	Many	Static	Zig-zag
s3.bag	Many	Dynamic	Straight
s4.bag	Many	Dynamic	Zig-zag
s5.bag	Not Many	Static	Straight
s6.bag	Not Many	Static	Zig-zag
s7.bag	Not Many	Dynamic	Straight
s8.bag	Not Many	Dynamic	Zig-zag

Sensor Fusion Combination List

• Wheel only
• Wheel + IMU
• LiDAR only
• Wheel + IMU + LiDAR
• Monocular
• RGBD Only
• LiDAR + RGBD
• Wheel + IMU + RGBD
• Wheel + IMU + LiDAR +RGBD

How to re-play rosbag dataset

• Download .bag file & copy to workspace

• Select scenario data by modifying the launch file.

  • Of course, rosbag options can be added through the "option" argument.

• View odometry visualization of sensor fusion method

roslaunch low_cost_odometry [sensor_fusion_method].launch

How to extract trajectory data

• Edit launch file

• Parameters

  • msg_type: argment name to select rosmsg type

  • extracted_topic: argment name to extract topic

  • scene_number: scenario folder name

  • tum_file_name: output tum file name

Installation EVO

pip install evo --upgrade --no-binary evo

plot output data

• Go to tum files directory

roscd low_cost_odometry/output/{s1, s2, s3 ...}
evo_traj tum wheel_only.txt wheel_imu.txt lidar_only.txt wheel_imu_lidar.txt rgbd.txt wheel_imu_rgbd.txt lidar_rgbd.txt wheel_imu_lidar_rgbd.txt --plot