Sprint 8: System Validation & Transition
Sprint Goal
Rigorously test the system to ensure accuracy, scalability, and reliability across all core features, and begin preparations for the final project presentation.
Sprint Planning
This final active development sprint focuses entirely on system validation and transitioning to the project presentation phase. We will conduct extensive scalability tests on the backend and evaluate the accuracy of our most complex detection algorithms (overtaking and safe entry/exit) to ensure the system is reliable for demonstration. We will also begin outlining the project video and poster.
Tasks
- Scalability Tests
- Overtaking detection accuracy evaluation
- Safe Entry/Exit detection accuracy evaluation
- Start planning the video and poster
Progress
- Ran scalability tests against the backend to understand how the message broker and detectors behave under increasing numbers of simultaneous vehicles, identifying the distributed-broker work needed to reach the 100,000-user target.
- Evaluated overtaking detection accuracy, including a real-car field test at the IT camp that confirmed the maneuver is correctly identified (surfacing a minor heading edge case when a vehicle is stopped).
- Evaluated safe entry/exit detection accuracy against varied map topologies and intersection types (as well as the IT camp field test), confirming the algorithm is robust across scenarios and identifying a minor edge case.
- Began planning the final video and poster, outlining the story, key demonstrations, and visuals for the project's presentation phase.
Challenges
- Generating realistic high-load scenarios to stress the backend without access to a large fleet of real vehicles.
- Measuring detection accuracy objectively across overtaking and safe entry/exit scenarios with limited ground-truth data.
- Reconciling the scalability findings with the non-functional target, which highlighted the need for a distributed broker architecture as future work.
Sprint Review
The final active development sprint closed the validation effort and prepared the project for transition. The core detection algorithms were confirmed reliable for demonstration, the scalability tests gave us an honest picture of the system's limits, and the team entered the transition phase with the video and poster planning underway.