Project 2 report is due 1 minute to midnight Sunday 29 November 2020 .
You can submit via WebCMS or by the terminal command
give cs3431 Project_2 zID_proj2.pdf.
Please make sure you include your zID in the file name as above and also remember to follow the template and out your zID, name and group name in your report.
In addition, one and only one team member should email to cs3431@cse.unsw.edu.au a zip or tar file containing all your source code.
The through report should probably be at least 5 pages, but no more than 10. As in project 1, make sure you are clear what your individual contribution is. It should be fairly detailed and it may be distributed across sections depending on what you contributed to the background, implementation and evaluation.
The project 1 marks have been entered into SMS finally.
The marks should now be visible in WebCMS.
Just reminder about MyExperience. I'd especially like some feedback on the group work. We've obviously had to work under unusually circumstances this year but I'd like to know how you think the groups have worked with our limited lab access. I hope we won't have to work under these conditions next time, but your comments would be very helpful for the future.
I've been able to book The Mech lab (J18) as well as the design studio (J17) for next Monday's practice session, so if you normally are in J18, please go there.
Yesterday I had to turn some people away because the design studio was becoming overcrowded, so please
email cs3431@cse.unsw.edu.au by Friday afternoon
if you are planning on coming in so we can manage numbers. If it looks like a lab will be overcrowded, we may have to limit teams to only two people.
If you are looking for possible 4th thesis projects for 2021, here is a set of projects related to our domestic robot. If you are interested, please contact us through cs3431@cse.unsw.edu.au.
If you are able to demo in week 10, please do, but since you haven't had as much time in the lab as previous years, due to COVID, we'll also allow demos in week 11. Please note that your individual report, is still due at the end of week 11.
Hello Everyone,
German has finished making a new simulation environment that is similar to the construct in the labs that you'll need to physically navigate with the Turtlebot. It has been pushed to the same git repository that you would have gotten the original model from at the following link:
https://github.com/stathiw/turtlebot3_gazebo
You should be able to pull from this repository since you should have already cloned it back in week one. Remember to recompile the code through catkin_make and source the setup.sh file afterwards or you won't be able to access it.
You can run it in the same way through the self_driving.launch file such as:
roslaunch turtlebot3_gazebo self_driving.launch
Please use this to test your code in simulation until you can access the labs during your tutorial time. Remember that if it works well in simulation however, there's no guarantee that it will also work well physically. Be sure to prioritise physical testing.
Colleagues in Biomedical Engineering are launching a startup and are looking for students who would like to do an industrial internship with them. The work is to write a ROS wrapper for their tactile sensor and to integrate in on an existing gripper. You can find information on here .
They already integrated their tactile sensors into an off-the-shelf gripper to demonstrate the advantages of the sensors being able to measure all of the tactile parameters that are important for gripping:
but they need to be ROS compatible to be useful in robotics.
To give you more time in the lab, instead of lectures on Monday we will be opening the J17 level 5 room for 12-3pm. Attendance is optional. We don't expect everyone to come to the Monday session. In fact , we can't fit everyone in, but if you feel you need the extra time, this will be available. There are a few constraints:
If you do plan to come to the lab on Monday, week 8, email cs3431@cse.unsw.edu.au first so that I can get an idea how many people are likely to turn up.
The week 7 lectures will be given by Dr. Wafa Johal. She will be talking about line recognition and tracking , so please attend the lecture as it is directly relevant to your project .
She will also talk about Human-Robot Interaction, which is a critical capability for robots, especially in home and office settings.
This is a reminder that the labs will be running this week. This will be your first chance to start working the self-driving car project. We'll set up some sample lanes to follow. Please also look at the autoqace code provided by Robots. This is not identical to the course we will use but it will give you some ideas about how to use OpenCV to do the vision and line finding.
The lecture next week will be given by Dr. Wafa Johal. She'll be talking more about computer vision, including recognising lines.
You can see a video taken from last year's final demonstration of one of the groups. The most important capability is lane keeping. They could also demonstrate stopping at intersections and recognising the coloured "traffic lights".
If you try to install the autorace demo, follow the instructions in the video. The web page says to install the 2020 version, but instead use the following git commands:
cd ~/catkin_ws/src/ git clone https://github.com/ROBOTIS-GIT/turtlebot3_autorace.git cd ~/catkin_ws && catkin_make
For anyone interested, after questions about the project, we will also talk a bit about the RoboCup@Home competition for anyone interested in joining the team. In RoboCup@Home, you get to work with this robot:
Week 6 is flexibility week, which means no new material is to be presented and no assessment. However, we will still have the labs so that we can get started on the second project. I'll also be online at the start of the Monday lecture period to answer questions about the project. How long I stay on depends on how many questions there are.
The second project is much more open-ended than the first one. We will not be giving very detailed specifications. The objective of the project is for us to see how creative you can be.
The minimum requirement will be that you can drive the TurtleBot3 around a driving course and stay within the lanes. You can improve the project by incorporating the following:
Tasks 3 and 4 may be doable in the time available. Tasks 1 and 2 may be achievable by a really good group.
To get an idea of how to go about this project, there are examples on the Robotis web site . There is a Gazebo simulator you can also try. Please note that this is not identical to our task. They use the Burger version of the TurtleBot3, whereas we have the Waffle Pi. Also, we will use dashed centre lines, not a yellow line. We will lay out some sample lane markings in the labs.
Note that the Gazebo simulation provided by Robots is unstable. It seems to work on some installations and fails on others. We are still investigating this, but this isn't our software, so we don't give any guarantees.
Try following this YouTube video (turn the sound down!).
The report for the first project should answer the questions in the assignment spec. The report template we provide is more intended for the final report.
You've already started some robot vision with the ARtags in Project 1 and this week's lecture introduce more robot vision, that you will need to know a bit about for the second project.
Before we start, we'll also answer questions on ROS transforms, which have been giving some trouble in the project. I've up some slides that give a general introduction and can go through those, but it may be more useful for you if you come prepared with any questions you might have on implementation details in ROS.
You can now submit your project reports through WebCMS or by "give cs3431 Project_1 report.pdf". You can chose a different report name, but it must be a PDF file with the ".pdf' suffice.
Reports are due 1 minute to midnight Friday 16 October.
Please let me know if you have any problems.
Getting the demos done by week 4 looks like it was too ambitious. Groups will be able to do their demos next week (week 5) and report's, which are individual , will be due on Friday week 5.
I just updated the assignment spec. I'd left in, from last year, mention of a group report. This time, only the individual report is required.
Just a reminder that Monday is a public holiday, so no lecture, but the labs go ahead as usual
Could we ask each group to note the name of their robot either as an additional team member in the group or if that's not possible (you may need a student ID to create a new member) then just add the name in as a comment in your specific group chat. Whatever robot you've got now you'll be using for the rest of the year so we need to make a note of who's got what
In case you missed these instructions, here's how you setup and launch the maze simulation in Gazebo.
Download updated Gazebo code:
Run wall follower code:
The wall following code may need to be tweaked to ensure smooth, collision free exploration of the maze.
Make sure you source your workspace each time you open a new terminal window.
ie. source ~/catkin_ws/devel/setup.bash
I think today's labs went reasonably well, but if you have suggestions for improvement please let us know. We have had suggestions for opening up the lab at other times, but unfortunately that is not an option due to Covid restrictions.
If you are working remotely, I please stay in touch with your team members in the lab, using Collaborate or any other tool you like.
Since we can get there people per group into the labs, one suggestion for next week is that one person who was present this week comes back with two new people so that the more experienced person can train up the others.
If there is a good reason for you to stay remote, please make sure that you are contributing to the group.
All the labs from now on will be for you to do your project work, so we won't be giving you much structure, but we will be here to support you. We'll have one tutor in each of the labs and one online to answer questions from remote participants.
I've done my best to assign people to groups. Please check the links below to make sure you are in the right group.
I may have got things confused where people wanted to switch between Tuesday and Wednesday, so please make sure that I got it right or let me know what to fix.
There are still some groups that have a mix of members on Tuesdays and Wednesdays. For now, I went with the day of the majority of members. We'll have to look at this again too.
With this arrangement, we may be able to get three members on a lab at one time.
In the lab this week, the groups will be assigned robots and you should start familiarising yourselves with the robot's operations.
If you are not in the lab, you should download the starter code for Project 1, compile it and start working on the project. You can learn more about running the SLAM (simultaneous localisation and mapping) nodes by following the TurtleBot tutorials that you started last week. We'll be covering these more in the lectures next week.
Thanks for forming up groups. We might have to make a few adjustments.
Running the labs this year will be less than ideal because we have to limit the number of students in each lab session, but we still want to have physical labs.
We'll discuss this more in the lecture.
Please make sure that you maintain proper physical distancing in the labs, even between team members. We will provide disinfectant wipes to keep the robots and your work areas clean.
So far there's only one group entered in WebCMS. If you are having trouble finding team members, we might form up groups for you, at least for the first project, so we can get started with the labs. If you have a group, but haven't entered it yet, please do that as soon as you can.
Hopefully, we will be able to also use the Design Studio for labs (J17-504), so we can get better access to the robots for you.
I've posted a draft specification for the first project under "Assignments/Projects". It may look complicated, but it is mostly taking off-the-shelf ROS packages and working out how to use them.
Because we have to work around COVID-19 restrictions, I'm prepared to flexible about submission dates are requirements, but note that the sooner we can finish this project, the sooner we can get started on the main project.
I've created two groups for Projects 1 and 2. You can keep the same group for both, if you want. If you have found your team members, please create a new group under the "Groups" tab on the left. Even if you don't have five, go ahead and create the group because then we can see how we can distribute the rest.
You should be able to access the recording of todays lecture by going to Moodle and Blackboard Collaborate. The link for the recordings is in the panel on the left (click on the three bars at the top left, if you can't see the panel).
The week 1 lab is online. You can find what to do here .
A Blackboard Collaborate session has been created so that we can talk you through the exercises and the tutors will be available to answer your questions as you work your way through the exercises.
Please note that in the original welcome message I had the lecture time incorrectly stated as 2-4. It is 12-2, as in the official timetable.
Please login to Moodle and select the Blackboard Collaborate link to access the online lecture.
Welcome to Robot Software Architectures.
Because of COVID19, we will run our lectures online, using Blackboard Collaborate . If this link doesn't let you in, you can access it via Moodle. This will be the only thing we will use Moodle for. The lectures will also be recorded. The first lecture is Monday 14 September 12-2pm. Please use the link above to attend the lecture.
Normally, we run this as a practical course, where students work in teams of 4 or 5, sharing one robot. This going to tricky in the current environment. The physical lab space available to us in J18-212. This is in the Willis Annex, next door to J17 and K17. To minimise physical contact, we will try to do as much preparatory work in simulation, but because we still want you to get practical experience with a real robot, groups may appoint a "robot wrangler" to run you experiments in the lab, so we reduce the number of student in the lab at any one time. We also suggest rotating who works on the robot, so that everyone gets a chance to experience the real thing.
This is the first time we've tried running this course online, so please be patient with us as we work out the best way of managing things. Your help and suggestions are welcome. We'll discuss how we'll run the labs more in the first lecture.
Please consult the course outline for details The assignments and projects pages are yet to be updated for 2020, but they gave you an idea of what to expect.