Local AI assistant for classic cars with voice control and vehicle integration
Transform your classic car with modern AI technology while preserving its authentic character. This system provides hands-free voice control for vehicle systems using a Raspberry Pi 5, local LLM processing, and direct vehicle integration via OBD-II and CAN bus protocols.
- Wake Word Detection: "Hey Car", "Vehicle Assistant", "Emergency Override"
- Natural Language: Conversational commands like "Make it cooler" or "Turn up the heat to 75"
- Multi-Turn Conversations: Context-aware interactions that remember your preferences
- Real-Time Processing: <600ms from voice command to vehicle action
"Hey Car, set temperature to 72 degrees"
"Turn on the air conditioning"
"What's my engine temperature?"
"Increase boost pressure by 2 PSI"
"Show me this week's performance summary"
- Privacy First: All processing on-device, no cloud dependency
- Ollama Integration: Local Llama 3.1 8B model with automotive training
- Hailo Acceleration: Dedicated AI chip for real-time inference
- Intent Recognition: Understands automotive context and vehicle-specific commands
- OBD-II Support: Read diagnostics from any 1996+ vehicle
- CAN Bus Control: Direct control of vehicle systems (make/model specific)
- Real-Time Monitoring: Engine temp, RPM, boost pressure, fuel economy
- Performance Analytics: Comprehensive data logging and trend analysis
- Multi-Layer Validation: Intent → Safety → Vehicle → Confirmation pipeline
- Emergency Protocols: Immediate safety override with "Emergency Override" command
- Driver Control: Manual controls always functional, system cannot override critical safety systems
- Safety Boundaries: Clear separation between controllable and critical systems
- Dual-Zone HVAC: Independent driver and passenger temperature control
- Smart Auto Mode: Learns preferences and adjusts automatically
- Voice Commands: "Set temperature to 72", "Turn on defrost", "Auto mode at 75 degrees"
- Safety Limits: Temperature and fan speed restrictions for comfort and safety
- Real-Time Dashboard: Web interface at localhost:8080 with live vehicle data
- Session Tracking: Automatic driving session detection and analysis
- Data Logging: Configurable detail levels from minimal to diagnostic
- Trend Analysis: Historical performance trends and efficiency monitoring
- Export Capabilities: CSV, JSON, and SQLite formats for external analysis
- Performance Tuning: Boost pressure, timing, and fuel system adjustments
- Real-Time Monitoring: Oil pressure, engine load, intake temperatures
- Safety Validation: All modifications validated against safety limits
- Diagnostic Integration: Read and clear trouble codes via voice commands
Temperature control (16-32°C) Fan speed (8 levels)
Auto/Heat/Cool/Defrost modes Air distribution control
Dual-zone support Automatic adjustments
Boost pressure monitoring RPM and load tracking
Oil pressure alerts Fuel economy analysis
Temperature management Diagnostic code reading
Interior lighting control Volume and source control
Brightness adjustment Playback management
Emergency flashers Speed-compensated volume
Comprehensive diagnostics Real-time performance data
Alert management Historical trend analysis
Export and reporting Maintenance scheduling
- Raspberry Pi 5 (16GB) with Samsung T9 SSD (2TB)
- 10.1" Capacitive Touch Screen for Raspberry Pi
- Hailo-8 AI Accelerator for neural processing
- 4-Microphone Array for 360° voice capture
- Vehicle Integration via OBD-II and CAN bus
- Automotive Power Management with 12V integration
- Voice Processing: Whisper STT, Piper TTS, Porcupine wake words
- AI Engine: Local Ollama with Llama 3.1 8B automotive model
- Vehicle Protocols: python-obd, python-can with safety validation
- Analytics: SQLite database with FastAPI web dashboard
- Safety System: Multi-layer validation with emergency protocols
Response Time: <600ms total RAM Usage: ~2GB with models loaded
Voice Accuracy: >95% clean CPU Usage: <30% during processing
Voice Accuracy: >85% road noise Storage: ~10GB system + models
System Uptime: >99.5% target Power: 12V automotive integration
# Clone and setup
git clone https://github.com/yourusername/automotive-llm-system.git
cd automotive-llm-system
python3 -m venv venv && source venv/bin/activate
pip install -r requirements.txt
# Run with full simulation
export AUTOMOTIVE_LLM_MOCK_MODE=true
python src/main.py --debug
# Try voice commands:
# "Hey Car, what's my engine temperature?"
# "Set temperature to 72 degrees"# Enable web dashboard
export AUTOMOTIVE_LLM_ENABLE_DASHBOARD=true
python src/main.py --debug
# Open browser to http://localhost:8080
# View real-time performance, manage alerts, export data# Clone and test immediately
git clone https://github.com/yourusername/automotive-llm-system.git
cd automotive-llm-system && python3 -m venv venv
source venv/bin/activate && pip install -r requirements.txt
# Quick verification (2 minutes)
python tests/test_system.py
# See the full test suite below ⬇️# On Raspberry Pi 5 with Raspberry Pi OS
sudo apt update && apt install -y python3-pip portaudio19-dev can-utils
git clone https://github.com/yourusername/automotive-llm-system.git
cd automotive-llm-system && python3 -m venv venv
source venv/bin/activate && pip install -r requirements.txt
# Install Ollama and AI model
curl -fsSL https://ollama.ai/install.sh | sh
ollama pull llama3.1:8b-instruct-q4_K_M
# Configure for your vehicle and run
cp config.yaml config-local.yaml
# Edit config-local.yaml for your vehicle
python src/main.py --config config-local.yaml- System Status: Live monitoring of all components
- Performance Metrics: Engine, thermal, and efficiency data
- Active Alerts: Real-time safety and performance warnings
- Session Tracking: Automatic driving session analysis
Live dashboard showing real-time vehicle performance metrics, system status, and analytics
- Configurable Logging: Minimal to diagnostic detail levels
- Multiple Formats: CSV, JSON, binary, and SQLite export
- Automatic Compression: Efficient storage with retention management
- Historical Analysis: Performance trends and comparative analysis
"Show me this week's performance summary"
"Start data logging for this trip"
"Export my driving data from last month"
"Do I have any performance alerts?"
"What's my average fuel economy?"
- Multi-Stage Validation: Every command validated through safety pipeline
- Emergency Override: Immediate system shutdown with voice command
- Parameter Limits: Configurable safety thresholds for all systems
- Fail-Safe Defaults: System defaults to safe state on any failure
- Driver Authority: Manual controls always functional
- Local Processing: No cloud dependencies or data transmission
- Encrypted Storage: All sensitive data encrypted at rest
- Access Control: Multi-level authentication and authorization
- Audit Logging: Complete command and action history
- Physical Security: Tamper detection and secure enclosure
✅ HVAC and climate control ✅ Engine performance monitoring
✅ Interior/exterior lighting ✅ Audio and infotainment
✅ Performance parameter tuning ✅ Diagnostic code management
❌ Braking systems ❌ Steering control
❌ Airbag systems ❌ Critical safety systems
❌ Transmission shifting ❌ Suspension control
# Basic system functionality (2 minutes)
python tests/test_system.py
# Full integration with data export (5 minutes)
python tests/comprehensive_test.py
# Live dashboard with real-time monitoring
python tests/enhanced_dashboard.py
# → Open browser to http://localhost:8080
# Realistic vehicle simulation
python tests/virtual_drive_test.py
# Run all tests
python tests/run_test_suite.py- 🎭 Mock Mode: Complete system simulation without vehicle hardware
- 📊 Live Dashboard: Real-time monitoring, debugging, and data visualization
- 💾 Data Export: CSV, JSON, SQLite formats with realistic automotive data
- 🚗 Virtual Drive: Multi-scenario vehicle simulation (city, highway, mountain)
- 📈 Analytics: Performance metrics, trending, and session tracking
📋 Complete Testing Documentation
📚 Complete Documentation Index
- Complete voice processing pipeline with wake word detection
- Local LLM integration with automotive-specific training
- Comprehensive vehicle interface (OBD-II + CAN bus)
- Full HVAC control system with safety validation
- Real-time performance monitoring and analytics
- Web dashboard with live data and controls
- Multi-layer safety system with emergency protocols
- Estimated Cost: ~$720-$750 for complete system (including 10.1" touch screen)
- Vehicle Compatibility: Any vehicle with OBD-II port (1996+)
- Connection Options: USB, Bluetooth, hardwired UART, or direct CAN bus
- Installation: Professional installation recommended for permanent setups
- Power: 12V automotive integration with backup systems
Over 50 voice commands implemented across climate control, engine management, lighting, audio, and analytics. Natural language processing with context awareness and multi-turn conversations.
We welcome contributions from automotive enthusiasts, developers, and classic car owners:
- Code Contributions: New features, bug fixes, optimizations
- Documentation: Guides, tutorials, vehicle-specific configs
- Testing: Hardware testing, vehicle compatibility
- Community: Share setups, help other enthusiasts
- GitHub Issues: Bug reports and feature requests
- GitHub Discussions: Questions and community support
- Documentation: Comprehensive guides and references
Important: This system is designed for classic cars with aftermarket modifications. Always prioritize safety:
- Driver maintains ultimate control of vehicle
- System cannot override critical safety systems
- Modifications may affect vehicle warranty
- Comply with local regulations regarding vehicle modifications
- Professional installation recommended for complex integrations
- Thorough testing required before road use
This project is licensed under the MIT License - see the LICENSE file for details.
⚡ Built for automotive enthusiasts who want to bring their classic cars into the AI age while preserving their character and charm.
🚀 Get Started Now | 📊 View Current Status | 📚 Browse Documentation