Week 11: The Synthesis , From Prototype to Proven System

As I reach the final full week of this 12-week project, the "noise" of building and coding has finally settled into the clarity of a finished system. A.E.G.I.S. is no longer a collection of sensors and jumper wires; it is a validated "Invisible Guardian" sitting in its final housing.

This week was dedicated to the big picture: summarising what the project actually achieved, reflecting on the steep learning curve of the past three months, and polishing the final academic documents, the 7,000-word dissertation and the 10-slide presentation.

Project Outcomes: What A.E.G.I.S. Achieved

The primary goal was to create a fall detection system that doesn't need cameras or wearables. After the intense testing in Weeks 9 and 10, I can officially state that the system is a fully functional bimodal prototype.

Final System Readiness:

Accuracy: The system achieved a 96% success rate in distinguishing between actual falls and daily activities (like sitting down or dropping objects).
Latency: The "Wave-Shift" engine responds in less than 450ms, ensuring that an alert is triggered almost instantly upon impact.
Stability: During my 8-hour stress test, the Raspberry Pi 5 handled the high-speed data from both the Wi-Fi and Radar threads without any crashes or thermal throttling.
Implementation: The prototype is ready for a "Pilot Study" in a real living room environment to see how it handles long-term ambient movement.

Knowledge and Skills Gained

Looking back at Week 1, my growth as an engineer has been massive, both in technical depth and professional mindset.

Mastering the "Physics of Radio": Before this project, I thought Wi-Fi was just for the internet. This journey forced me to view it as a high-precision sensor. I had to master:

CSI Processing: Learning how to extract 64 subcarriers and perform NumPy matrix math in real-time.
Sensor Fusion: Solving the problem of two sensors "disagreeing" and building a weighted logic gate to find the truth.
Embedded Optimisation: Managing serial buffers and CPU cycles on the Pi 5 to ensure the code never "stutters."

The Engineering Mindset: I learned that in engineering, a "bug" is just a data point. When the radar data was lagging in Week 8, I didn't see it as a failure; I saw it as a requirement for a "Buffer Flush" function. I've learned to document every mistake, as those are often the most important parts of the final report.

Evaluating Impact: The Privacy Victory

A.E.G.I.S. solves the "Privacy vs. Safety" dilemma I identified in my Literature Review.

The Dignity Factor: The system achieved its goal of 100% camera-free operation. This means elderly users can be protected in private areas like bathrooms without feeling watched.
Autonomy: By solving the "Quiet Room" problem with the UDP ping, I ensured the system stays active even when no one is moving.
Passive Protection: Unlike a pendant, the user doesn't have to "do" anything. The system just exists in the background, which is a huge win for user compliance.

Limitations and Future Upgrades

An honest engineer has to admit where the system can be improved. A.E.G.I.S. is a successful prototype, but it has boundaries.

Current Limitations:

Single-Room Scope: The current setup covers one medium-sized room. For a whole house, I would need a mesh network of ESP32 nodes.
Large Pet Interference: Very large dogs might cause enough wave-scattering to confuse the Wi-Fi engine, though the radar's human-height logic helps filter this.
Manual Start: The system still requires me to run a script manually. It isn't yet a "plug-and-play" consumer device.

Proposed Improvements:

Mesh Expansion: Integrating multiple ESP32 beacons to cover an entire flat or house.
Emergency API: Using a service like Twilio to automatically send a text message or place a phone call to a caregiver when the "Red Alert" is triggered.
Auto-Boot: Setting up the Python script as a "System Service" so it starts automatically when the Pi 5 is plugged in.

Report and Presentation Status

Dissertation Final Polish: I am currently in the final stages of the Results and Discussion chapter. I am making sure that every hardware photo I took matches the data in my testing logs. My goal is to make the report read like a professional engineering audit.

Presentation Prep: My 10-minute viva presentation is ready. I’ve structured it to show the Problem (Privacy), the Solution (Wi-Fi + Radar), and the Proof (the 100% confidence captures). I am practicing my demo to make sure I can explain the "Confidence Weighting" clearly to the examiners.

Final Reflection

This 12-week journey has changed how I look at technology. A.E.G.I.S. proves that with smart "Hardware-Software Co-Design," we can build life-changing tools that respect human privacy.

Next Week: The Final Submission. The 12-week marathon is over.