Embedded Systems
FlagshipSmart Solar Tracking System with Adaptive Energy Management and Weather Protection
A solar panel that follows the sun, protects itself from bad weather, forecasts solar behavior, and makes its performance visible.
Context: Final bachelor thesis / engineering capstone
Period: 2024-04 to 2025-05
Summary
This thesis joined a physical solar-tracking installation with local monitoring and forecasting. The panel could orient itself toward stronger light, report electrical and environmental telemetry, react to rain or vibration, and expose its operation through a web dashboard running from a Raspberry Pi.
My role
Individual bachelor thesis: I designed and integrated the embedded control, local platform, telemetry, dashboard, energy logic, and forecasting workflow.Scope and next validation step
Testing covers the working thesis prototype. A longer seasonal deployment would be the next step for evaluating outdoor durability and forecast behavior across a full year.2017–2018
NSRDB training period
2019
GHI prediction period
5 days
Recent history for next-day operation
10
Meteorological input features
From sunlight to a useful decision
Panel and sensors
Light, power, battery, rain, vibration, and weather readings
ESP32
Real-time sensing, tracking, actuator control, and protection
Raspberry Pi 5
Local orchestration, processing, and forecasting
Django and PostgreSQL
APIs, telemetry history, and remote actions
Dashboard
Live state, trends, weather context, and forecast output
What the system handled
Directional photoresistors and dual-axis servos moved the panel toward the strongest available light.
Voltage, current, battery, temperature, humidity, rain, vibration, and light readings formed one telemetry stream.
Rain and impact events could trigger protective behavior automatically.
Historical weather data and recent local readings supported pre-installation assessment and next-day forecasting.
Technology
Physical control
Local platform
Monitoring
Forecasting
Conclusions
- The result was one working prototype that connected the hardware, dashboard, and machine-learning components.
- Keeping real-time control on the ESP32 and orchestration on the Raspberry Pi gave each layer a clear responsibility.
- The forecasting work connected historical NSRDB evaluation with a practical next-day mode designed around recent local readings.
- Outdoor power, noisy sensors, servo movement, persistence, and dashboard responsiveness had to be considered together.
Keep exploring