1Pose Graph 2D 2---------------- 3 4The Simultaneous Localization and Mapping (SLAM) problem consists of building a 5map of an unknown environment while simultaneously localizing against this 6map. The main difficulty of this problem stems from not having any additional 7external aiding information such as GPS. SLAM has been considered one of the 8fundamental challenges of robotics. A pose graph optimization problem is one 9example of a SLAM problem. 10 11This package defines the necessary Ceres cost functions needed to model the 122-dimensional pose graph optimization problem as well as a binary to build and 13solve the problem. The cost functions are shown for instruction purposes and can 14be speed up by using analytical derivatives which take longer to implement. 15 16Running 17----------- 18This package includes an executable `pose_graph_2d` that will read a problem 19definition file. This executable can work with any 2D problem definition that 20uses the g2o format. It would be relatively straightforward to implement a new 21reader for a different format such as TORO or others. `pose_graph_2d` will print 22the Ceres solver full summary and then output to disk the original and optimized 23poses (`poses_original.txt` and `poses_optimized.txt`, respectively) of the 24robot in the following format: 25 26``` 27pose_id x y yaw_radians 28pose_id x y yaw_radians 29pose_id x y yaw_radians 30... 31``` 32 33where `pose_id` is the corresponding integer ID from the file definition. Note, 34the file will be sorted in ascending order for the `pose_id`. 35 36The executable `pose_graph_2d` has one flag `--input` which is the path to the 37problem definition. To run the executable, 38 39``` 40/path/to/bin/pose_graph_2d --input /path/to/dataset/dataset.g2o 41``` 42 43A python script is provided to visualize the resulting output files. 44``` 45/path/to/repo/examples/slam/pose_graph_2d/plot_results.py --optimized_poses ./poses_optimized.txt --initial_poses ./poses_original.txt 46``` 47