Report.txt

### Team Information:
-> 10569363, Luca Alessandrini
-> 10571436, Mattia Portanti
-> 10829293, Massimo Buono


### Files Description
In our src folder appears three packages:
1) localization: this package contains two nodes: tf_publisher.cpp, which is the node requested to publish the odometry as tf, and trajectory.cpp, the node needed to create the trajectory which is published on topic path, used for visualization.
Moreover this package contains the two launch files needed to run the nodes and publish the static tf.
1.a) mapping.launch: it is used to start gmapping, what we use to create the map, with all the parameters. To merge the scans, we included the launch file "laserscan_multi_merger.launch" present in the package ira_laser_tools.
1.b) amcl.launch: it is used to start the localization with amcl. For simplicity, we wrote all the parameters of amcl in a separated file, amcl.yaml (config folder of localization package). In this launch file, we also call our custom node in the package map_server, "map_saver_trajectory", needed to save the pgm file of the trajectory.
2) ira_laser_tools: this is the package used to merge the scans. We modified the launchfile laserscan_multi_merger.launch by putting the correct values of: 
	->topic names
	->minimum angle
	->maximum angle
Moreover we computed an increment angle such that we did not exceed the maximum number of scans supported by gmapping.
3) map_server: this package, cloned from github, contains an additional custom node "map_saver_trajectory.cpp". It is inspired from "map_saver.cpp", and it can be used to save the .pgm file of the map with the trajectory. 
This save the map with trajectory in .ros folder interpolating the points from /amcl_pose topic.(clearly, it is an interpolation, so we have no assurance that the robot is passed from there, but it is reasonable to assume that, since we need to have a continuous movement, since the velocities are not extremely high, and since the time between two pose messages is small).


### TF structure
map -> odom -> base_footprint -> base_link
                                        |_ laser_front
                                        |_ laser_rear

### Bag usage
-> robotics1_final.bag -> map creation
-> robotics2_final.bag -> localization test
-> robotics3_final.bag -> localization test


### Node used for map creation
To create the map, we used slam_gmapping of gmapping package.
To save the map created we use map_saver of map_server package.


### How to start and use
To create the map:
	-> roslaunch localization mapping.launch
	-> rviz in order to visualize the creation (we included our config file in the submission)
	-> rosbag play [path/to/bags]/robotics1_final.bag --clock
	-> save the map with: rosrun map_server map_saver -f map
		-> Notice that that the map must have name "map"
		-> Notice that with this command the map is saved in the folder where the command is called
			-> if not already done, put the created map.pgm and map.yaml files in the "maps" folder of localization package

To test the localization:
	-> be sure to have the map.pgm and map.yaml files in the "maps" folder of the localization package.
	-> roslaunch localization amcl.launch
	-> rviz in order to visualize the creation (we included our config file in the submission)
	-> rosbag play [path/to/bags]/robotics2_final.bag --clock (or robotics3_final.bag)
	-> to save the trajectory: rosservice call /SaveTrajectory [file_name]
		->Notice that the saved trajectory pgm and yaml file will be placed in the .ros folder of the computer.


### Important/Interesting things
In order to save the trajectory, we re-wrote an existing node. We added a service to it, and, considering the map as a matrix, we performed the interpolation between each two subsequent points. In order to print the trajectory, since the saved file is in grey-scale, we used an intermediate value between occupied and free in order to make it visible.