MediLocate is an indoor navigation system designed for hospitals where GPS signals are unreliable. It utilizes WiFi Signal Strength (RSSI) fingerprinting combined with Machine Learning to accurately determine a user's location (e.g., Specific Lab, Hallway, Doctor's Room).
The system consists of an ESP8266 based signal scanner, a Python-based ML Engine for classification, and a Cross-Platform Flutter App to visualize the user's position on the hospital map.
- Data Collection (Offline Phase):
- Using ESP8266/Mobile, we scan WiFi Access Points (APs) at known locations.
- A dataset is built mapping RSSI vectors to specific zones (e.g.,
Lab 1,Corridor).
- Model Training:
- A Classification Model (Random Forest / KNN) is trained on the RSSI dataset to recognize the unique "fingerprint" of each location.
- Real-time Localization (Online Phase):
- The app scans current WiFi signals.
- The model predicts the location based on live signal strength.
- The result is plotted on the digital floor map.
- Mobile App: Flutter (Android/iOS/Web).
- IoT/Embedded: Arduino C++ (ESP8266 for signal acquisition).
- Machine Learning: Python (Scikit-Learn, Pandas).
- Data: CSV datasets of signal strengths.
The model is trained to recognize specific zones within the facility:
- Labs: (Lab 1, Lab 2, etc.)
- Corridors: (Hall Right, Hall Left)
- Offices: (Doctors' Room)