Weekly Blogs

Most of this week, I've been spending on trying to make the duckiebots work. Despite previous assumption that all the bots work, it seems only horizon and skyduck are functioning right now. One of the biggest revelation for me so far has been, the bots starts malfunctioning when they're running low on battery. It requires a lot of time to charge them fully and it drains out even faster. So far, I've been able to make skyduck turn on and set up the accounts required to move it around. However due to the lengthy time it takes for flashing the SD card and charging the bots, my plans have been hampered quite a lot. My main goal for the next week is to find out proper ways to charge the duckiebots through the slack and stack overflow channel and ensure it's fully charged so that I can progress with my work properly.

Here's the list of task and updates that I've worked on this week:

Progress:

• Went through the duckiebot manual operations to complete all the operations tasks given in the website.

• Created a ROS Project for operating the duckiebot using dockker images

• Among all the operations: making it move was the hardest to control for some issues that were debugged.

• Created a subscriber and publisher file for the DT project for making the duckiebot move.

• Able to control the duckibot using keyboard controls. The wheels are calibrated and able to do line following for some time(skyduck)

• Able to run the gym-stimulation as well

Issues:

• The keyboard control for skyduck was not working on my computer or Anna's computer. The issue lied on having a heavy RAM disk. We had to manually clean up space to free some disk space and were able to control the robot again using the keyboard.

Here's the list of task and updates that I've worked on this week:

Progress:

• Got introduced to coding with C++. It's a little bit difficult at first to get accustomed  with but it will be okay hopefully.

• Created and set up PPL on the remote ubuntu desktop

• Installed vizmo and played around it!

• Evaluated different MP strategies by playing around the k-numbers and different MP strategy

• Read the motion planning paper. I think the paper is great resource for someonw to understand the underlying work in motion planning and does a great job of explaining all the research that has been going on it for the past decades

Issues:

• I had a little bit issue creating the alias for the first task but other than that nothing major happened in terms of the 5 tasks given.

• C++ is a little bit difficult to understand.

• The paper in general was written using pretty much simple languages. However, for some who doesn't have any idea about the terms or languages related with Motion planning, I had difficult time trying to understand at time what he was trying to deliver. It took a lot of googling and figuring out to understand the meassage he was trying to deliver but I got through it.

Here's the list of task and updates that I've worked on this week:

Progress:

• Read thoroughly the code for collision validity checker in an attempt to understand how they check for collision.

• Studied the lane following code and went deep into reading the image projection, line filter and all the launch files related with lane following.

• Checked the publisher and subscriber code base of lane following for better understanding.

Issues:

• I think I've a  good idea of the structure of the code that I studied but I would have liked if I had more guidance on how I can walk towards more implementation on how it will be impactful or resourceful for our own research.

Here's the list of task and updates that I've worked on this week:

Progress:

• We have created the google sheet containing all the data for all the strategies in  both 3D and ZTunnel(except for DR-PRM)

• Query and Lazy-Query for all 3 have been done usin 15 different seeds and we have analyzed the data for it except for DR-PRM so far.

• Tried rebooting skyduck, ellenduck and leafduck. 

• 
  

Issues:

• Ever since we reflashed the SD card for skyduck, we are unable to find it usin fleet discover. So we have troublesoot that. Disassembled ellenduck in an attempt to assemble it and reboot it again.

Here's the list of task and updates that I've worked on this week:

Progress:

• Solved all the issues related with DR-PRM. There are still a number of underlying issues that needs to be fixed but for now it's working.

• Ran all the strategies in 3D environment and calculated the all the necessary metrics in data sheet.

• Learned about mining environment.

• Worked on the poster for the symposium and finished it. 

Issues:

• Need to study DR-PRM deeply to understand the issue that is causing it to fail and successfully run it again.