Assignment 1
DEADLINE: 23rd December at 23:59 (CET)
Implement a routine that lets Tiago navigate inside the environment, see Figure 1. The
environment comprises two rooms with movable obstacles (i.e. you can move them in
Gazebo) and a narrow space. Tiago has to navigate from point Starting Pose to Pose_B
(Figure 1).
Figure 1
The Pose_B must be inserted by the user via the command line, i.e., it is an input from the
command line for your code. Once the robot reaches Pose_B, it has to detect the obstacles,
and print on the screen the position of the movable obstacles, e.g. the cylindrical tables,
visible from that pose. (Figure 2). The static obstacles that are part of the map, e.g. the
walls, or the shelves, must not be detected as obstacles.
Figure 2
Suggestion: use the laser range finder mounted on the robot (topic /scan).
3
move_base result status I
se
Action È
Result list ofpositions geometry_msgsPositioned
For the assignment, you have to create a package which contains the code to implement.
Your code must be implemented using the following structure:
● The user can input the pose of Pose_B by command line;
● Your code must be implemented with an Action Client/Server structure:
○ The action client receives the input from the user;
○ The action client calls the action server that executes all the tasks;
○ The action server sends the final list of positions of the obstacles as result to
the action client;
○ The action client also implements callbacks to the feedback of the action
server, and prints the task's current status in the terminal. The feedback must
reflect the current status of the robot, e.g.: the robot is stopped, the robot is
moving, the robot started the detection of the obstacles, the detection is
finished, etc.
Extra points
Typically, the ROS Navigation Stack is inefficient in narrow passages. Usually, the
programmer has to implement an ad hoc routine to directly process the sensor data (e.g.,
laser and/or sonar data) and generate velocity commands for the robot, i.e. the programmer
creates a “motion control law” for the robot. In this assignment, to get extra points, we ask
you to implement your control law to navigate through the narrow corridor using the laser
data. The routine has to use the laser and send the velocity commands to the topic /cmd_vel
to move the robot without calling the move_base stack.
N.B.: Please, specify in the submission in the report that you implemented the extra point
part of the assignment.
Instruction to start the homework
Local ROS Installation
Who has a local ROS installation (e.g. Ubuntu or Virtual Machine) is required to install the
Tiago packages. Please, follow these instructions.
Install the Tiago simulation workspace:
> sudo apt-get update
> sudo apt-get install git wget ipython3 python3-rosinstall
ros-noetic-desktop-full python3-catkin-tools python3-rosdep
python-is-python3 ros-noetic-actionlib-tools
ros-noetic-moveit-commander
> sudo apt install -y ros-noetic-four-wheel-steering-msgs
ros-noetic-urdf-geometry-parser ros-noetic-ddynamic-reconfigure
ros-noetic-people-msgs ros-noetic-twist-mux ros-noetic-map-server
ros-noetic-amcl ros-noetic-robot-pose-ekf
ros-noetic-moveit-planners-ompl
ros-noetic-moveit-simple-controller-manager ros-noetic-global-planner
python3-future ros-noetic-teb-local-planner ros-noetic-hokuyo3d
ros-noetic-urg-node ros-noetic-sick-scan
> mkdir ~/tiago_public_ws && cd ~/tiago_public_ws
> wget
[Link]
devel/tiago_public-[Link]
> source /opt/ros/noetic/[Link]
> rosinstall src /opt/ros/noetic tiago_public-[Link]
> catkin build
> source src/[Link]
> rosdep install -y --from-paths src --ignore-src --rosdistro noetic
--skip-keys "urdf_test omni_drive_controller orocos_kdl pal_filters
libgazebo9-dev pal_usb_utils speed_limit_node camera_calibration_files
pal_moveit_plugins pal_startup_msgs pal_local_joint_control
pal_pcl_points_throttle_and_filter current_limit_controller
hokuyo_node dynamixel_cpp pal_moveit_capabilities pal_pcl
dynamic_footprint gravity_compensation_controller pal-orbbec-openni2
pal_loc_measure pal_map_manager ydlidar_ros_driver"
> cd ..
Create a workspace to develop your packages
> mkdir catkin_ws && cd catkin_ws
> source ~/tiago_public_ws/devel/[Link]
> mkdir src && cd src && catkin_init_workspace
> cd .. && catkin build
Now you are ready to install the environment for the assignment inside catkin_ws
● Clone your repository into your workspace, it will initially contain all the environment
and launch files necessary to start the development. You must create your package
into your group repository and deliver this one as a whole:
> git clone
[Link]
● To remap to a new repository it can be helpful remove the old git blob:
> rm -rf tiago_iaslab_simulation/.git
● Build the package:
> cd ~/catkin_ws
> catkin build
> source ~/catkin_ws/devel/[Link]
● Start the simulation (cmd console 1):
> roslaunch tiago_iaslab_simulation start_simulation.launch
world_name:=robotics_library
● Navigation stack (cmd console 2):
If I enable 30 acceleration higherFPSbutlaserdisappears
> roslaunch tiago_iaslab_simulation [Link]
N.B. depending on your GPU setup, you may face some problems in the laser simulation. Anyway,
you can use VLAB to correctly simulate laser sensors.
VLAB
In VLAB, the tiago environment is already installed inside the container.
● As usual, please follow the instructions to start the container:
> start_tiago
> source /opt/ros/noetic/[Link]
> source /tiago_public_ws/devel/[Link]
● Create a new workspace (you can also use one already created):
> mkdir -p ~/catkin_ws/src
> cd ~/catkin_ws
> catkin build
> source ~/catkin_ws/devel/[Link]
● Download the package with the simulated environment into the workspace:
> cd ~/catkin_ws/src
> git clone
[Link]
You are not supposed to edit this package. Just download and use it.
● To remap to a new repository it can be helpful remove the old git blob:
> rm -rf tiago_iaslab_simulation/.git
● Build the package:
> catkin build
> source ~/catkin_ws/devel/[Link]
● Start the simulation (cmd console 1) and wait until the start up is completed:
> roslaunch tiago_iaslab_simulation start_simulation.launch
world_name:=robotics_library
● Navigation stack (cmd console 2):
> roslaunch tiago_iaslab_simulation [Link]
N.B. Sometimes the simulation does not properly initialize. If you experience some issues,
try to kill and restart the last two commands.
How to submit your solution
To submit your solution, you must do the following:
1. Setup your group repository
a. Create your private group repository on Bitbucket named as
“ir2324_group_XX”, where XX must be substituted with your group number:
> cd ~/catkin_ws/src
> git init
> touch [Link]
> git add [Link]
> git commit -m “first commit”
> git push
If you are using local installation, please be sure to upload in the repository only the
packages you are required to submit, for example calling:
> echo -e ‘tiago_iaslab_simulation/’ >> .gitignore
b. Includes all the group members and the 3 tutors in the collaborators of the
repository. Here our BitBucket profiles:
● [Link]@[Link]
● [Link].1@[Link]
● danieleva2005@[Link]
c. Public repositories or repositories we cannot access will not be considered
for correction.
2. Start coding
a. Push your code (with good comments) into your group’s git repository.
Code explainability will be part of the evaluation criteria.
b. We encourage you to commit the code as you write it and not in one block
c. Add a [Link] file which contains the necessary commands to
v correctly execute and test your code (e.g. roslaunch and rosrun commands).
Make sure the instructions are working properly. The first lines of the
[Link] must contains the following informations:
GROUP XX
Member1’s name, email
Member2’s name, email
Member3’s name, email (if any)
d. We suggest you to test your code step by step in a new clear workspace,
cloning the code from your repository in order to be sure that your code is
executable by third parties.
e. Make sure your code is compiling before submitting. Not compiling codes
will be penalized in the final grade. We will make a small compilation test a
week before the deadline in order to let you know if your code compiles
properly. We will post on the forum the test modalities.
3. Submission
1. You are allowed to push the code in the repo until the submission on
Moodle. Later commits will not be considered for the evaluation.
2. To get the maximum grade, you have to submit within the 23rd December at
23:59 (CET). Later submissions will be penalized.
3. Prepare a brief PDF report (max. 2-3 pages) that explains your solution.
The report is supposed to explain the high-level ideas, strategies or
algorithms you implemented to solve the assignment. The first lines of the
report must be like:
GROUP XX
Member1’s name, email
Member2’s name, email
Member3’s name, email (if any)
LINK TO THE REPOSITORY
LINK TO ADDITIONAL MATERIAL (see point 5)
4. Prepare a short video (e.g., screen recording) that shows the execution of
your software by the robot. This is necessary in case we cannot easily run
your code.
5. Report and video must not be placed in the repository. We suggest you
create a Google Drive folder, upload the video or any additional material and
share with us the link to the folder inside the report.
6. We will open a submission to Moodle. You will be able to attach the report to
your submission (be sure that it contains all the needed information such as a
link to the repo, link to the video, additional info, etc.). The name of the
submission must be as follows: GroupXX_A1. Please submit only once
per group, one member can submit for everybody.
LaserScan 1 ApplyDBSCANtofindclusters
2 Runcirclefittingalgorithmtoeachcluster
anglemin 110º
angle max 110º ya i
10º
i sia
size int
m nt
o i increment anglemin
The DBSCAN algorithm seems to be working decently, however the circles found by the
hyper tting algorithm are just wrong.
TODO:
Extract results of the DBSCAN algorithm (the single clusters) and apply the algorithm
to those clusters.
Extrapolate and adapt the tting algorithm in a standalone main to test this.
Visualise the results.
0.15 20.20
Filter theresultsbyradius
MATTEI
mobile_base controllercondvel
1 Understand how tocontroltherobotusingendvel
2 Createsimplefunctiontosend geometry_msgs Twist linearandangularvelocities
goal
walls
Keepthedistancefrombothwalls
thecontrollawsseenin RIC2 workfortheidealmodel
adapttodifferentialdrive doneimplicitlybytheTwist
g
robot 7 thesetwogiveusthewidthofthecorridor
rare parallel
Estimations [Link] re Eiea fEn Y
and distancesfromthewall
091 TL 4wg99fyheckleft right
suppose we'realignedwith thewall 90 and 90 measurementsgive usthewidthof
thecorridor but
Option B defineandtracka trajectory thiswouldsolvethe problem is oneshot
Issues howdo yougenerate the trajectory butwehaveto understandhowto generate
atrajectoryusinglaserscandata
0 rridor
howdo 1,4a tk
91choose
you waypoint
motwaY
1 Obtainthewidthofthecorridorbylookingat 90º thecorridorwidthistheminimumsumof
the 2 distances
2 Define waypoint as final position at theend of thecorridor based on distancesat 0º
3 Controllawtoreachwaypoint
4 Control tomaximisedistancefromwalls
