Overview
Remote outdoor environments and off-grid facilities often lack reliable internet connectivity, rendering cloud-dependent smart assistants (Alexa, Google Assistant) completely non-functional.
The Solara V-CORE project delivers an autonomous, green-energy solar lighting station that responds to spoken natural language commands locally and instantaneously without needing WiFi or cellular networks.
Core Hardware Architecture
- VC-02 Offline Speech Processing Module: Embedded neural acoustic processor trained on localized trigger keywords ("Turn On Light", "Eco Mode", "Dim 50%", "Solar Battery Status"). Achieves greater than 96% recognition accuracy within 5 meters with sub-100ms response time.
- Main Actuation Controller: Arduino microcontroller managing PWM constant-current LED drivers, relay safety gates, and power routing.
- Solar Energy Harvesting Subsystem:
- Monocrystalline photovoltaic solar panel with MPPT charging circuit.
- LiFePO4 battery pack monitored via dedicated ADC voltage dividers and current shunt resistors.
- Ambient Light LDR Sensor: Prevents daylight illumination waste through automatic twilight threshold gating.
Key Features & Capabilities
- Zero Cloud Dependence: All speech feature extraction, neural model matching, and hardware actuation happen entirely on edge silicon.
- Auditory Feedback: Integrated speaker delivers localized Vietnamese/English voice confirmations upon executing commands.
- Low-Power Deep Sleep Mode: System idles at under 15mA while maintaining active acoustic wake-word listening.
Source Code & Schematics
Full firmware and acoustic command model weights are available on GitHub - Solara-Voice-System.