yuwati’s summer blog: week_7_progress

day_5_july26

Progress Today

  • Worked on Planning II videos and completed them. The main topic was Graph Theory.
  • Did not finish the Planning I task but discussed it with Danielle. Danielle had finished the task.

Tentative Plans for Next Week:

  1. We will get to know/choose our individual tasks next week.
  2. It is a good idea to do the section on Estimation and Localization and Reinforcement Learning. They are a part of planning motions for the robot.
  3. Finish going through the developer manual.
  4. Work on the Planning Task –> may give me an idea on how we can implement algorithms with duckiebots. It looks like duckiebots use Python.

day_4_july25

Goals For Today

  1. Work on the Planning I & II sections of the Edx course and complete their exercises
  2. Continue working on the Duckietown Developer manual (will save for tomorrow or next week)

Progress

  • Finished going through the quizes and the videos for Planning 1. The main topics were about Formulizing Motion Planning Algorithms, Two types of motion planning: Path Planning and Kinodynamic Planning, and Dimensionality, Workspace, and Configuration Space.
    – Currently working on the Collision-Checker exercise forked from the duckietown-lx repository.

day_3_jul24

Goals for Today

  • I submitted a pull request to gym-duckietown.
    I made changes to the gym_duckietown/simulator.py with an updated version of a function that was causing trouble. circle ci test was failing, but this could be another function depreciation. Reason: Docker images have pre-installed? versions. This one looks like it is running a python3.7. Python 3.7 does not support numpy 2.0.0. This could be the problem.
  • We did a paper review.
    Danielle and I reviewed a paper on the Moving Visibility Graph with Obstacles – Travelling Salesman Algorithm. The paper proposed a complete solution to the traveling sales man problem with moving targets and stationary obstacles.

day_2_jul23

Goals for Today:

  1. Continue working on the developer manual
  2. Docker alternative of the simulation?

Progress

  • [ERROR] noticed today that leafduck did not shutdown yesterday.
    I ran a DT project on leafduck and it doesn’t seem to be able to shut down since then.
    Since this morning, I have done a dts duckiebot shutdown leafduck and manually clicked on the top button for 5 seconds, but the duckiebot remained on.
    [FIX?] I tried I did dts duckiebot reboot leafduck and it did not shutdown, it lost connection with dts fleet discover. After waiting for a few minutes, I manually pressed the top button, and it shutdown. I am not sure if it can successfully turn back on yet. It’s LED reaction (turning blue) when its charged seems to be working fine.
  • Working on the Docker alternative to setup the simulation
    1. Liam Paul from Duckietown got back to me on Slack and suggested that I try out the docker alternative.
    nvidia-docker pull duckietown/gym-duckietown && \
    nvidia-docker run -it duckietown/gym-duckietown bash
    but I ran into an error. Error: nvidia-docker command not found
    2. I remember that I tried to install nvidia-docker but it did not work out. I followed a suggestion from chat gpt and successfully installed the docker image mentioned on the daffy branch of gym-duckietown repository.
    3. Ended up switching to branch master.
    4. However I ran into another error. Because the dependencies and the files were on the docker image, there was an error (there was no error message) in opening up the GUI for the simulator. It did not open, but I could see the updates of the coordinates? in the terminal. The xfvg viewer had probably been deprecated. Some of the recommended lines of code in the guide could not be run, because the terminal could not recognize the commands. More information in the rose_lab documentation.
    More info in this Slack thread: https://duckietown.slack.com/archives/C012QLMA5MZ/p1720816002605449
  • Completed fixing the simulation error.
    – I abandoned the docker way to run the simulation and went back to

day_1_jul22

Goals for today:

  1. Work on developer manual
  2. work on simulation

Progress

  • Working on the developer manual:
    While working on updating and pushing code to the duckiebot, I ran into some errors with numpy dependencies. The most up-to-date numpy is 2.0, I used 1.26, but the duckiebot is only compatible with numpy 1.24.4 at the latest.
    – Completed Beginner DTProject tutorial.
    Everything went well here. I did not run into any errors. I think it will be a good idea to write down a personal manual on the steps to take to make a DT project
    – Starting on Beginner Use ROS project (on step: ROS publisher)

Leave a Reply

Your email address will not be published.

Name *
Email *
Website