12/7/2024
This week, I finished integrating the GPS with the radar code, as well as with the Traffic Light Detection code. I implemented polling with multithreading for both. I also integrated Lane Detection code with Traffic Light Detection code, implementing the logic for determining the corresponding traffic light relative to the users’ position. I also built the track that our team is using for the demo, replicating roads and lanes mostly to scale to our remote control dump truck. I also did integration testing for combining the TLD and LD code, and helped retrieve testing footage from the demo environment.
My progress is on schedule this week. I have completed a lot of the integration of our features as well as adding new features. I am still testing them, but next week I will start gathering testing metrics.
Next week, I hope to finalize the software portion of the project, as well as get testing metrics for both our final report, demo, and poster. Our team will be designing the poster, editing our final video, and writing our final report in hopes to complete the project. We must finalize building the track, and I will be adding on an audio cue to alert the user when necessary.
11/30/2024
This week, I finalized the purchases for the battery pack and demo material, making adjustments based on what power source we need for our new demo. I did research on how many volts the Jetson Orin Nano requires and how long we would need power for, confirming the research I did previously. I also found different models of RC cars that would best fit the metrics of our system, as well as fit our budget. The traffic light detection algorithm was not detecting color correctly, and I had to do tests and debug on different videos as to why it wasn’t working. I realized the bounding box of where I was detecting light intensity was not aligned with where the bounding boxes of the traffic lights were. I now have the basic algorithm of the TLD finalized, and now just need to finetune based off our entire system. In preparation for the final presentation, I created the skeleton for the final presentation as well as editing the schedule for the next weeks for our final sprint.
My current progress is as up-to-date as possible. A lot of my progress is now dependent on other portions of the project, like integrating GPS and lane detection as well as testing for the final demo. Our team hopes to complete those dependencies early next week, so I hope to make my final changes by the end of the next week.
In this next week, I hope to be done with most of the integration of the entire system, as well as testing it on our demo environment. This will allow us to do integration testing, and create the alerting system for the user. While there have been lots of adjustments to our project, we are hopeful to have a working demo by the end of the semester.
11/16/2024
This week, I created new versions of traffic light detection and started testing them on pictures after the main object detection from our YOLOv8 model. I tested some images, making sure that our latest version based on light intensity position was able to correctly identify whether or not the light was red or green. I also began testing the lane detection, but realized that the lanes were not being properly detected. I’m currently working on improving the lane position, but also focusing on getting images detected and classified for the interim demo. In addition to this task, I helped Emily with her task of finding a powering system for the Jetson Orin. I found forums that talked about powering their Jetson with a laptop charger that connects to a car smoker. I also then found the portable battery source that we could use for the RC cars, as well as a way to use the same cable of connecting a car smoker to the DC battery source.
My progress this week has been on track with the traffic light detection, and it was nice to see my traffic light detection code be able to detect lights. However, I will need to continue to debug lane detection and help the team combine all of our components, as well as continue testing.
Next week, I hope to have a running version of lane detection and combine the lane detection, traffic light detection, and machine learning components. I will also be focusing on our interim demo, making sure we have enough to present and share.
11/9/2024
This week, I worked on a lot of the traffic light detection and lane detection code. I looked into examples on Github as well as other sources to code, using OpenCV and Python to detect traffic lights with HoughCircles and Hough Transformations. I created two versions of the traffic light detection code, one including a YOLOv8 model to detect the bounding boxes of traffic lights, and the other using HoughCircles to detect green and red circles in the frame. This is because our team was unsure of if the YOLOv8 model would finish training in time, so we wanted to have a backup plan to make sure we could make progress. After multiple iterations of developing this code, we came to the conclusion that tracking frames and traffic light positions is only necessary when the car is not moving, so incorporating the GPS will be our next step. Because of this change, I had to reimplement some of the logic with this code. I also tried to do initial development of our mobile application, but due to equipment issues and limitations, we had to change our mobile application development tech stack to React Native instead of Swift.
My progress is behind schedule because training the YOLOv8 model turned out to take longer than anticipated. In addition to this, including outside elements like the GPS and the lane position has not been completed. However, I will have adequate time to speed up the testing and development process.
Next week, my main focus is the mobile application and finishing the initial development of the TLD and LD. After I finish these, Ankit will connect these elements with the other remaining parts of the project and the Jetson.
11/2/2024
This week, I focused and led our team on ordering our new equipment, making sure the new camera and GPS fit our requirements of price, plug and play abilities (streaming), and compatibility with our Jetson Nano. I also began some trial code for traffic light detection as well as color changes on the traffic light. Although we don’t have our camera yet and do not know how we will detect traffic lights, we believed it would be beneficial to start some of the programming to detect color as this doesn’t solely depend on the camera.
Since our materials have been reordered, our project schedule has shifted. Emily has adjusted our project schedule to include these new changes. This next week will be focused on developing and synchronizing our 3 pieces of data: camera, GPS, and radar, as well as finish the initial development of traffic light detection and lane detection.
10/26/2024
This week, I worked on initial setups and downloading dependencies for the Jetson Orin. After the Jetson stopped working, I also helped our team with troubleshooting. I created a Github repository for our team to collaborate with version control, and I also started the data extraction process from the dashcam.
My progress is currently behind schedule, as I have not been able to work on the lane detection feature. After figuring out running code on the Jetson as well as extracting data, this process will be able to be completed. I will continue to work on this feature next week, and will have enough time to complete the other features at the end of the week. Our team has ample slack time, so I believe my progress will not alter our final product.
Next week, I hope to finish the Lane Detection feature as well as get started on the traffic light detection feature.
10/20/2024
This week, I helped my team complete the Design Report – mostly focusing on the Use-Case Requirements, Architecture and Principle of Design, and Design Requirements. I also helped Ankit with his sections, specifically with information regarding the mobile application that is part of our project. Our original design requirements were not clearly outlined from our previous presentation, so I spent most of my time ensuring that our design requirements were fully fleshed out.
My schedule is on track with our schedule, but I personally was not able to work on any of my stretch goals. I hope to be able to contribute more to the mobile app in the next few days.
Next week, I will help the team figure out how the data will be transferred and used from our sensors and cameras to our hardware, as well as get basic measurements for latencies to keep track of progress in regards to our design report. I will also look into more traffic light detection functionalities.
10/5/2024
This week, I prepared for and gave our design presentation. I practiced presenting and was able to give the presentation successfully. I also started working on the Design Report as well as some of the mockups for the mobile application. I set up the mobile application development portion, and am currently waiting on our parts to be delivered.
My progress is on schedule, as most of our development is scheduled for the next 6 weeks.
Next week, I’m hoping to create some buttons and pages for the mobile application on Figma and to ensure that our Jetson is working properly with Ankit. I also plan on ordering the rest of our material so it is ready for pickup by the end of fall break. I also plan on finishing up my section of the Design Report and overall editing and reviewing the entire paper.
9/28/2024
Throughout this week, I focused on equipment procurement and ordering as well as the design presentation slides. I was able to find a radar sensor that meets our needs and requirements while being within our budget. I ordered the camera, radar, and radar enclosure. I focused on creating a block diagram with system specifications, making sure that important details were on the slides. This ensures that connections between hardware are clear, as well as the function of each block in the diagram. I also researched useful OpenCV examples that relate to our project, as well as other documentation and helpful notes that will be used later during the implementation of our software (opencv-lane-detection, cow-detection-opencv, etc).
My progress is on schedule.
By next week, I hope to order the rest of the materials we need for MVP and to finish the Design Report with the rest of our team. As a continuation of my stretch goal from last week, I hope to create a high-fidelity wireframe of the mobile application.
9/21/2024
This week, I fine tuned the proposal presentation slides and created a quick wireframe of the mobile application of our product, AutoAlert. I created some User Interface (UI) elements, including checkboxes and a slider. I also organized notes and clarified the individual homework each team member should complete by our next meeting. I read through the design presentation guidelines, creating rough ideas for each section of the design presentation. I also did research on what exact products we are interested in, including the dash camera, wire connections, and sensors.
My progress is currently on schedule. Next week, I hope to finish up the design presentation slides and practice presenting the slides by Thursday. I also hope to finalize what equipment to purchase as my first task is to complete the equipment procurement. As a stretch goal, I hope to create a high-fidelity prototype of the mobile application, as well as begin to create a rough skeleton of the application using Swift.