Find parking. Avoid tickets. Stay informed.
Parkopticon is a cross-platform mobile app that helps drivers find open street parking and avoid parking enforcement through crowdsourced, real-time user reports.
Parkopticon empowers drivers with community-driven intelligence about:
🅿️ Available Parking Spots - Real-time reports of open street parking- 👮 Enforcement Alerts - Warnings when parking officers are nearby
- 🚗 Parked Car Protection - Get notified if enforcement approaches your vehicle
- 🎫 Ticket Management - Log tickets and get tips for appealing them
This repository contains only the web/mobile application for Parkopticon. The edge inference and autonomous reporting system is being developed in a separate repository.
⭐ Smart Vehicle Sentinel & Enforcement Detection System — Project Summary (Resume-Ready)
I designed and implemented an AI-powered, low-power vehicle sentinel system that uses edge computing, mmWave radar, distributed IoT nodes, and cloud-based deep learning to detect nearby enforcement vehicles and pedestrian activity while maintaining strict power efficiency constraints for overnight operation.
The system integrates embedded hardware, computer vision, wireless networking, and microcontroller-driven event orchestration to provide real-time situational awareness around a parked vehicle.
⭐ Key Responsibilities & Technical Contributions • Edge AI & Computer Vision Pipeline
Built an on-device inference pipeline using an Orange Pi 5 (RK3588S) with a 6-TOPS NPU, deploying quantized YOLO-based object detectors and MobileNet classifiers for vehicle and pedestrian identification.
Implemented a two-stage detection system combining NanoDet/YOLOv5n on edge and high-accuracy YOLOv8-L/Vision Transformer models in the cloud to classify enforcement vehicles with sub-0.1% false-negative rate.
Optimized inference for 1–3 frame wake events, reducing compute cost by >85% while maintaining accuracy.
• Embedded Systems & IoT Node Design
Architected a distributed mesh of ESP32-S3 microcontroller nodes equipped with:
mmWave radar for motion and vehicle signature detection
Ultra-low-power accelerometer (LIS3DH) for tamper and vibration sensing
Camera modules with rolling video buffers stored on microSD
Designed these nodes to operate at 0.3–0.6W, enabling continuous monitoring without draining the vehicle battery.
• Low-Power System Engineering
Engineered the core system to remain under a 50 Wh/12-hour power cap, using:
Suspend-to-RAM modes on the SBC
Interrupt-driven wake-on-GPIO
mmWave-triggered selective inference
ESP32-based rolling dashcam buffers to avoid keeping the SBC active
Modeled and validated energy usage across multiple operating states (idle, inference bursts, LTE upload).
• Sensor Fusion & Distributed Event Processing
Implemented a multi-sensor event pipeline combining:
mmWave radar signals
Motion classification
Vehicle silhouettes
Accelerometer tamper events
GPS/geolocation metadata
The system intelligently wakes the central compute module only when necessary, reducing noise and power consumption via edge-filtered events.
• Connectivity & Cloud Integration
Integrated LTE Cat-1/Cat-4 modem modules (SIMCom A7670/7600 series) with Canadian band support for secure image uploads and cloud verification.
Implemented a cloud inference API for high-precision confirmation of enforcement vehicle detections, transmitting only cropped frames for bandwidth efficiency.
Built fallback mechanisms to operate offline with local-only inference.
• Real-Time Analytics & Alerting
Developed logging, timestamping, sensor fusion, and event categorization for:
Enforcement vehicle detection
Pedestrian proximity alerts
Vehicle movement around the car
Impact and tampering events
Designed an extensible structure for future push notifications / mobile app integration.
⭐ Technologies Used
Edge AI: YOLOv5/8, NanoDet, MobileNetV3, RKNN Toolkit Embedded Systems: ESP32-S3, mmWave radar, accelerometers, OV2640/OV5640 cameras, microSD storage SBC & OS: Orange Pi 5, RK3588S NPU, Armbian/Linux, GPIO wake Networking: LTE Cat-1/Cat-4 modules (SIMCom), UART/USB modems, ESP-NOW, Wi-Fi Low-Power Design: Suspend-to-RAM, duty cycling, wake-on-interrupt, power budgeting Cloud: ONNX Runtime, YOLOv8-L, API-driven verification pipeline Other: GPS (u-blox NEO-8M), distributed IoT event orchestration, system-level integration
⭐ Impact
Engineered a system that provides 360° real-time sentry monitoring with <50 Wh overnight power usage.
Achieved high-precision enforcement vehicle detection using a hybrid edge/cloud inference architecture.
Built a novel distributed IoT + edge AI solution with applications in vehicle security, parking analytics, and privacy-centric autonomous monitoring.
It will connect to the app and act as an autonomous reporting node.
- Interactive Map View - See available spots and enforcement in real-time
- Report Parking Spots - Share open spaces with optional photos
- Report Enforcement - Alert others to officer sightings, chalking, or ticketing
- Smart Notifications - Get push alerts if enforcement is near your parked car
- Ticket Logger - Track your tickets and learn how to appeal them
- Cross-Platform - Works on both Android and iOS
- Real-Time Updates - Live map with crowdsourced data
- GPS Integration - Accurate location-based reporting
- Photo Upload - S3-style cloud storage for spot photos
- Push Notifications - Background location monitoring for alerts
- Node.js (v20+)
- Android Studio (for Android emulation)
- VS Code (recommended)
- Expo Go app (for testing on real devices)
# Clone or navigate to the project
cd "k:\Self Improvement\Coding\Park-Opticon\parkopticon"
# Install dependencies
npm install
# Start the development server
npx expo start
# Press 'a' for Android emulator or scan QR code with Expo GoSee docs/setup/SETUP_GUIDE.md for complete installation instructions including:
- Figma design setup
- Android/iOS emulation
- VS Code configuration
- Troubleshooting tips
For all documentation, see docs/README.md
parkopticon/
├── App.js # Main app entry point
├── app.json # Expo configuration
├── package.json # Dependencies
│
├── assets/ # Images, icons, fonts
│ ├── icon.png
│ ├── splash.png
│ └── adaptive-icon.png
│
└── src/ # Source code
├── theme/ # Design system (colors, typography, spacing)
├── screens/ # App screens
├── components/ # Reusable UI components
├── navigation/ # Navigation setup
├── services/ # API & external services
└── utils/ # Helper functions
See docs/development/PROJECT_STRUCTURE.md for detailed explanation.
- React Native - Cross-platform mobile framework
- Expo - Development tooling and managed workflow
- React Navigation - Screen navigation
- React Native Maps - Interactive map interface
- React Native Paper - Material Design components
- Expo Location - GPS and geolocation
- Expo Camera - Photo capture
- Expo Notifications - Push notifications
- AsyncStorage - Local data persistence
- FastAPI - Python REST API
- PostGIS - Geospatial database
- PostgreSQL - Data storage
- AWS S3 - Photo storage
Phase 1: Frontend Development Environment ✅
- Project setup with Expo
- Basic map view with markers
- Theme system (colors, typography, spacing)
- Location permissions
- Interactive reporting (tap to add markers)
- Development documentation
Phase 2: Core Features (In Progress)
- Navigation system (Bottom tabs)
- Report screens (Parking spots & Enforcement)
- Camera integration for photos
- Local data storage
- Notification system
Phase 3: Backend Integration (Planned)
- FastAPI backend server
- PostGIS geospatial queries
- Real-time data synchronization
- User authentication
- Cloud photo storage
Phase 4: Advanced Features (Planned)
- User profiles and reputation system
- Historical data and patterns
- Smart notifications (ML-based)
- Ticket appeal guidance
- Community moderation
Design files are maintained in Figma. The app follows:
- 8pt grid system for consistent spacing
- Material Design principles for familiarity
- Accessibility standards for inclusive design
- Primary: #2196F3 (Blue) - Trust and reliability
- Success: #4CAF50 (Green) - Available parking
- Error: #F44336 (Red) - Enforcement alerts
- Warning: #FFC107 (Yellow) - Caution
npx expo start
# Press 'a' to launch on Android- Install Expo Go from App Store/Play Store
- Start dev server:
npx expo start - Scan QR code with Expo Go
- Android Emulator: Use "Extended Controls" → Location to set GPS coordinates
- Real Device: Enable location services and grant permissions
Complete documentation: docs/README.md
- docs/setup/SETUP_GUIDE.md - Complete environment setup
- docs/setup/QUICK_START.md - 5-minute quick setup
- docs/design/DESIGN_QUICK_REF.md - Design overview
- docs/development/PROJECT_STRUCTURE.md - Code organization
- docs/setup/CHECKLIST.md - Setup verification
This is currently a solo project, but suggestions and ideas are welcome!
- Design screens in Figma
- Implement UI in React Native
- Test on emulator and real device
- Iterate based on feedback
"Cannot connect to Metro bundler"
npx expo start -c # Clear cache"ADB not recognized"
- Add Android SDK to PATH (see SETUP_GUIDE.md)
Emulator is slow
- Install HAXM (Intel) or enable Hyper-V
- Allocate more RAM in AVD settings
See docs/setup/SETUP_GUIDE.md Section 8 for more troubleshooting tips.
This project is for educational and personal use.
- Expo for excellent development tools
- React Native Maps for map integration
- Material Design for UI guidelines
- The open-source community for amazing libraries
Questions or suggestions? Open an issue or reach out!
Happy coding! Let's make parking easier for everyone. 🚗
