This week I worked on developing a ROS node to subscribe two topics which supply LIDAR data as well as the location data from the camera which tells the ROS node where the car is located in the frame. I worked on integration between the two systems. Next week, now that the camera is working, I will be testing combining the data from the two systems.
Fayyaz Status Report for 4/16/22
This past week I helped two main things. The first being helping Ethan with the Lidar. We were able to use the TimeSynchronizer package that synchronizes incoming channels by the timestamps contained in their headers to allow for clear data input from both the Lidar and the camera at the same time. We used this website as a reference (http://wiki.ros.org/message_filters) .
The second thing that I worked on was finally working on the physical design aspect of the system. I thought this would be done the best through using SolidWorks and using 3d printing. After measuring the length, width, and height of all the components, I sketched a 6x5x6.5 inch box with shelves to hold the components. Below is a picture. 
After submitting the design, though, I was told that the printing failed and the design was too big. From here, I decided to pivot to acrylic and wood design and put the pieces together.
Now that Chad got the camera working, I feel a bit better and believe that we are on schedule. In the next week, I hope to make the physical components with the acrylic, have the Jetsons talking to another (if Chad’s work allows for it), and start prepping for the final presentation.
Ethan’s Status Report for 4/10/22
Unfortunately I was sick for a few days this week so I was unable to make as much progress on the LIDAR system as I wanted to. After I recovered on Monday, I started writing a ROS node that will accept incoming LIDAR messages and use them in conjunction with the object recognition system to figure out distance to cars. I also tested the LIDAR and collected some range data as shown below. Next week I will be working with my other teammates to integrate the LIDAR and camera together. 
Team Status Report for 4/10/22
As a team, we are trying our hardest to ensure the delivery of our project. The most significant risk right now is the camera still not working. As we have reflashed the jetson and are using some new libraries, we hope it should be working soon but there are definitely some hardware limitations when it comes to running the YOLO algorithm. This is being managed as we are working on other aspects of the project and ensuring those are working such that when the camera is figured out, we can move swimmingly. If it does not work on the Jetson, we were thinking of switching to a RPi to see if the space and CPU there could handle what we have.
No changes as of right now to the design of the system.
We did update the Gaant chart and schedule to account for the delay with the camera.
Fayyaz Status Report for 4/10/22
This past week I was helping with a few things but due to carnival and the demo, nothing incredibly substantial. After the failures with the camera on one of the Jetsons, this week I worked on reflashing the device. This was different than previous as this time I was trying to see if there were settings or libraries install earlier in setup that could have contributed to the disfunction of the Yolo algorithm. After consideration, I decided to only flash and do the base setup to ensure nothing could be done. This was done and it seems there have made some progress with the new algorithm.
Besides this, I have written a basic script for the two jetsons to talk with one another and hope to implement it when both jetsons are done with their individual components.
My progress is a bit behind as it depends on both the camera and Lidar working which is not the case right now. As this is the case, I hope to come in and work with chad to see what we could to speed up the camera process and ensure there is something that we can work with soon.
In the next week, I hope to deliver the new script I created as well as a basic design (again) of our cage to house the components.
Chad’s Status Report for 04/10
For this past week, I spent more time trying to get the Yolo algorithm working on the Nano. Not much work was done because of carnival and also the demo on Monday. I re-flashed the SD card on the Nano to retry installing the necessary libraries for the YOLO algorithm so I will continue working on this for the coming week. The schedule is still slightly delayed because of YOLO being so difficult to install on the Jetson Nano. There is a chance I will have to use a different algorithm instead if it still doesn’t work after this week.
Fayyaz Status Report for 4/2
This past week I transitioned to working on two main components. The first being the design of the project. After looking at the individual designs of each of the different aspects of the project, I came to realization that maybe using two jetsons would be better than one. After discussing with the group, we thought this idea would be great. From here, I started looking to the physical components that could house all the parts of the project. I decided to looking in to some 3d modeling programs to design it.
More importantly, after deciding on using two Jetsons instead of one, I was trying to figure out how to have both jetsons talk to one another. After talking with the professor, he mentioned to use Ethernet. I did some research and found this website as a reference: https://youtu.be/VbuHzujRcNo . After the others are done setting up their work more and focusing on other aspects, I hope to focus on this.
I think my progress is on schedule. There is a bit more work now but, I think we are still on track.
Over the next week, I hope to implement the jetson communication and have some base physical design to present.
Team’s Status Report for 4/2/2022
This week we split up the work on our project. We worked on both getting the object detection algorithm running on one of the Nvidia Jetsons, and we worked on getting the LIDAR working on the other. Since the object detection algorithm is so heavyweight, we are considering the idea of using both Nvidia Jetsons. One will do the object detection and the other will combine data from the object detection with data from the LIDAR and display the output to the user’s glasses. Next week we will be working on integrating the two systems.
Ethan’s Status Report for 4/2/2022
This week I mostly worked on getting the LIDAR to work with ROS. I was able to install ROS on one of the two Nvidia Jetsons and write a ROS node that would read the Laser scan data from the LIDAR. Next week I will be working on starting to write the code which combines the information from the object detection and the LIDAR and displaying it to the HUD glasses.
Chad’s Status Report for 04/02
This past week, I was still having many issues with installing the Yolo Algorithm on Jetson Nano. I had installed the PyTorch and Torchvision libraries by following the steps on this website:
https://forums.developer.nvidia.com/t/pytorch-for-jetson-version-1-10-now-available/72048
At first there was some issues but I was able to solve it by downloading the arch version of anaconda on the Nano called archiconda3. This version included python v3.7 which I was able to use to complete the installation process listed above. However when trying to run the JetsonYolo.py from the JetsonYolo GitHub, or any python file that uses the OpenCV library for some reason doesn’t have permissions to use the CSI-camera module on the Nano. I keep receiving a “Unable to Open Camera”. After doing a deep dive, there was apparently an issue with the version of the OpenCV algorithm I was currently using so I am now installing one of the newer versions of the OpenCV package in hopes that it will fix this issue.
Its a very long installation process and while installing, warnings were popping up saying there may not be enough storage on the Jetson so this may pose an issue. There have been multiple problems showing up while trying to get Yolo installed on the Jetson so this has set back my personal schedule a bit. The goal for next week is that I can hopefully get the Yolo detection fully working on the Jetson.
