This week, I focused on designing the computer vision pipeline and refining the overall system architecture, with particular emphasis on adapting the solution to real-world constraints. Based on power considerations, I redefined the hardware requirements by transitioning the edge platform from an NVIDIA Jetson to a Raspberry Pi, ensuring better alignment with system limitations while maintaining functional goals. I also spent time finalizing key details of the project proposal, including design rationale and implementation considerations. Overall, my progress remains on schedule, and this hardware adjustment helps reduce risk moving forward. In the coming week, I plan to begin validating the CV pipeline on the updated hardware platform and support the team in completing remaining proposal deliverables.
Team Status Report for 2/14/2026
Overall Project Status
This week, the team made solid progress across hardware, software, and documentation by ordering key components (Jetson, gps transponder, power supply, and cellular note carrier) for the dash cam solution, prototyping the overall dash cam design, and establishing an initial backend architecture along with a well-organized codebase. We also defined several core design proposal components, including the design requirements, system specification and block diagram, implementation plan, and the test, verification, and validation plan, while aligning on project management responsibilities for the immediate next steps. At this stage, no major design changes were made, as the current system architecture meets the project requirements; the primary risks identified involve hardware integration and on-device performance, which are being mitigated through early prototyping and staged testing. The project remains on schedule. In the coming week, the team plans to finalize and order the remaining parts, set up the Jetson SDK, finalize the database schema, establish the Firebase backend, and configure a consistent developer workflow to support coordinated development moving forward.
Attribution:
Part A was written by Dylan.
Part B was written by Giovanni.
Part C was written by Kevin.
Part A: Public Health, Safety, and Welfare Considerations
PotSpot directly addresses public safety and welfare by improving the efficiency and reliability of pothole detection and reporting. Poor road conditions pose safety hazards to drivers, cyclists, and pedestrians, and manual reporting methods are often infrequent and delayed. By automatically detecting potholes with high accuracy and uploading precise location metadata, the system enables faster identification and repair of hazardous road conditions, reducing the likelihood of accidents and vehicle damage.
The system is designed for fully hands-free operation after initial setup, ensuring that drivers are not distracted while operating their vehicles. This design choice prioritizes both physical safety and driver well-being while maintaining continuous data collection in the background.
Part B: Social Factors
Social considerations for PotSpot center on trust, privacy, and equitable access to infrastructure improvement tools. While the system collects road imagery, it is designed to focus on pothole detection rather than individuals, and access to pothole status updates and management features is restricted to authorized city administrators through authentication and audit logging. This helps mitigate concerns related to surveillance and misuse of data.
Additionally, PotSpot supports collaboration between drivers, local governments, and communities by aggregating crowdsourced data into a centralized, easy-to-understand interface. This encourages civic engagement and improves communication between residents and municipalities regarding road maintenance priorities.
Part C: Economic Factors
Economically, PotSpot is designed to balance system performance with cost efficiency. Hardware components were selected to meet strict accuracy and latency requirements while maintaining reasonable weight and power constraints, keeping the overall device affordable and scalable. The use of cloud-based infrastructure enables the system to support up to 1,000 active users without requiring costly on-premise server deployments.
By enabling faster pothole detection and repair, the system has the potential to reduce long-term infrastructure maintenance costs and vehicle repair expenses for drivers. Clear, data-driven prioritization of road repairs allows municipalities to allocate resources more efficiently, resulting in economic benefits for both local governments and the broader community.
Dylan’s Status Report for 2/14/2026
This week, I personally made progress on several core technical and planning tasks, including prototyping the computer vision pipeline, developing an initial dash-cam–based system concept, and researching suitable computer vision datasets to support model training and evaluation. In parallel, I contributed to the design proposal by working on the proposal slides to clearly communicate system goals, constraints, and architecture. Overall, my progress is on schedule relative to the project timeline, with no major blockers identified this week. In the coming week, I plan to focus on on-device testing and iterative tinkering with the prototype CV pipeline to evaluate real-world performance, as well as finalize the remaining components of the parts list to prepare for hardware integration.
Dylan’s Status Report for 2/7/2026
I focused on building the foundation for the computer vision portion of the project by researching different YOLO versions and computer vision algorithms relevant to pothole detection, as well as exploring CV pipelining strategies for edge deployment. I also began assembling the parts list and submitted order forms for the Jetson Orin Nano, while ensuring steady progress against the project schedule. Next, I plan to finalize and complete the parts list, place the remaining orders, and start experimenting with training and testing computer vision models for pothole detection on the target hardware.
Dylan’s Status Report 1/31/2026
I primarily contributed to the proposal slides by refining use-case requirements and clearly defining project tasks. I also researched existing datasets and evaluated different approaches for pothole detection to inform our technical direction. Overall progress remains on schedule. Moving forward, the next steps include prototyping and testing multiple YOLO model versions, exploring suitable camera hardware modules, setting up an initial code versioning workflow using GitHub, and brainstorming ideas for the computer vision pipeline.