report

SECOND PROJECT ROBOTICS 

Gabriele De Santis 10820992
Alessandro Gaeta   10632035
Tommaso Pieroni    10801962



Small description of the files inside the archive:

Folder: <launch>
Has 3 files:

- amcl.launch.xml: is the launch file were we set the frames and the parameters for amcl.

- gmapping.launch: is the launch file used to launch gmapping. We found of particular relevance, in order to create a good map, properly setting some parameters such as the maxRange, maxUrange particles, linearUpdate and angularUpdate.   

- launcher_pro.launch: is the main launch file used to launch the map_server, tf, rviz and robot_localization nodes. In the latter one, we specified all the frames necessary for the localization. Here we also launch the yaml file with the map created with the bag 1.


Folder: <maps>

In the maps folder you can find the files map1.pgm and map1.yaml which we created using gmapping on the bag 1.bag.


Folder: <rviz>

In the rviz folder there is the config file rviz_config.rviz used to launch rviz with a proper configuration for the scenario we are working on. 

Folder: <tf_tree>

Contains an image of the working tf tree.



Bag used to create the map -> {1.bag}
Bag used for the tests	   -> {2.bag, 3.bag}


In order to launch gmapping ( to see the creation of the map ), inside a ROS environment type:
$ <roslaunch project_robotics gmapping.launch>.
Obviously make the selected bag (1.bag) play and then, to see the real-time creation of the map, open rviz and add the Topic "map".

In order to launch the localizations nodes, together with RVIZ, inside a ROS environment type:
$ <roslaunch project_robotics launcher_pro.launch>.
Obviosuly make the selected bag play.

In order to perform localization we chose the visual odometry + IMU (from which we get {X, Y, Yaw_dot} from the visual odometry, and {double derivative of X and Y} from IMU). We found that this sensors are the most reliable.
We fused to it the scout odmoetry from which we get {X, Y, Yaw}. We chose this sensor because we found that also this sensor was reliable enough.