This past week, I spent most of my time coalescing the datasets of the trash items we wish to detect into respective classes and also thinking/researching ways to further tune the YOLOv7 and YOLOv7-tiny models. As I’m worried about wasting compute Read more…
This past week, I have been working on setting up ROS on the Raspberry PI. I have also been working on learning ROS and defining motion requirements for the robot. This Monday, I delivered our team’s design review presentation. Overall, Read more…
The most significant risk right now is the usage of tiny Yolo V7. According to MATOM Ai, although tiny Yolo model processes faster and lightweight, it is less accurate than Yolo V7 (MATOM Ai, 2022). This means it can give Read more…
This week, there were many tasks I personally accomplished on the project. First, I worked on design presentation slides and then did some research on a camera module compatible with the Jetson Nano Orin Developer Kit as the following: https://docs.google.com/document/d/1qags4xzWa6l6wguN0HfArgNEiNxMWlKxdvZhU5k9h0o/edit?usp=sharing Read more…
This past week, I spent most of my time setting up Google Colab and training the YOLOv7 and YOLOv7-tiny models. As expected, the YOLOv7 has been tuning well, but the metrics for the YOLOv7-tiny model are not so stellar. Since Read more…
This past week, I spent a great deal of time working on the design for the robot. I researched existing trash collecting robots (on land/ water) and other mechanisms that collect objects (ex. vacuum cleaners, applications in farming, cleaning, toy-collecting Read more…
This week, I presented our project proposal. I also spent time further researching different pre-trained ML model options and looking into different datasets available for use. As a part of this process, I took a closer look at the existing Read more…
This week, I primarily researched different pick up mechanisms for our garbage cleaning robot in order to finalize our parts ordering list. Hirani and I are collaborating on a design sketch for the upcoming design presentation. Along with the process, Read more…
This week, I spent time researching components and defining design constraints for our design. In continuation of the use case requirements we defined in the proposal last week, I came up with several new design requirements that helped aid in Read more…
The most significant risks of the project right now whether the pick up mechanism, in coordination with the rest of the design, will work well and integration between the Jetson Nano Orin and Raspberry Pi. After thorough research, receiving feedback Read more…