AutoSense

AutoSense is an Edge-IoT vehicle monitoring project based on an ESP32-S3 OBD-II/CAN adapter, a React Native mobile app, Firebase services, and a LaTeX thesis document.

Repository: https://github.com/Fonio22/AutoSense

What It Does

• Reads vehicle telemetry from OBD-II over CAN.
• Stores compact binary history locally on the ESP32-S3.
• Streams live telemetry to the mobile app through BLE.
• Exports unsynced binary logs from the ESP32-S3 to the app without deleting them from flash.
• Uploads historical log files to Firebase Storage and stores sync metadata in Firestore.
• Uses lightweight edge analysis with statistical baselines, Z-score, and Tiny Isolation Forest.
• Documents the project in the thesis under Writting/main.tex.

Repository Layout

OBD2-Firmware/       ESP32-S3 firmware, OBD-II service, BLE protocol, logger, anomaly analysis
app/                 Expo / React Native mobile app and Firebase rules
vehicle-profiles/    Vehicle profile definitions, including vw_passat_2016
kiCad/               PCB design and schematic files
outputs/             Processed datasets and exported analysis inputs
Writting/            LaTeX thesis source, figures, references, and compiled PDF

Firmware

cd OBD2-Firmware
pio run

Main modules:

• src/app.cpp: boot sequence and main loop.
• src/obd_service.cpp: PID discovery, polling, and OBD-II decoding.
• src/obd_binary_logger.*: compact binary log storage and export.
• src/obd_ble_protocol.*: BLE commands, telemetry stream, and log sync.
• src/anomaly_detector.*: baseline, Z-score, Tiny Isolation Forest, and severity levels.

Mobile App

cd app
npm install
npx tsc --noEmit
npm run lint
npm run ios

The app uses Expo Router, Firebase Auth, Firestore, Firebase Storage, SecureStore, BLE, and local file storage for downloaded OBD logs.

Profile resolver tests:

cd app
node --test functions/test/profile-resolver.test.mjs

Thesis

cd Writting
latexmk -pdf -interaction=nonstopmode -file-line-error main.tex

Main files:

• Writting/main.tex: thesis source.
• Writting/main.pdf: generated thesis PDF.
• Writting/Reference/references.bib: bibliography.
• Writting/Figures/: diagrams, screenshots, test photos, and result charts.

Validation Snapshot

The current thesis documents:

• Volkswagen Passat 2016 as the main real-vehicle validation profile.
• OBD-II simulator testing with a generic profile.
• A 60-day route-classified dataset with city, highway, and unknown segments.
• Functional tests for acquisition, BLE/USB communication, storage, visualization, and analysis.

Safety Scope

AutoSense is designed as a read-only monitoring system. It does not erase diagnostic trouble codes, does not write to vehicle ECUs, and does not replace professional mechanical diagnosis.