SolarCore

Initial Infrastructure

Hardware Initialization

ESP32 Firmware Flash

Upload the core firmware bundle to the ESP32. Ensure the following pins are correctly mapped in your config.h:

MPU6050 (I2C): SDA (21), SCL (22)
LDR Array: A0, A3, A4, A5 (Analog)
Motor Control: PWM Pins 18, 19, 23 (Azimuth/Elevation)

Raspberry Pi Setup

For AI-ANN enabled hubs, install the SolarCore Linux environment:

                            sudo apt update && sudo apt install python3-pip flask

                            pip install tensorflow-lite numpy requests

Network Provisioning

Configure LAN/Wi-Fi credentials. The system will attempt connection on boot. Fallback: If connection fails, the hub enters Offline Mode utilizing local LDR/RTC data.

Hardware Infrastructure Registry

Core variables and system logic parameters tracked within the hardware layer:

MPU Current Angle

Bi-directional XYZ vector tracking. Supports both hardware feedback (output) and manual angle injection (input) for remote operation.

LDR 1-4 Intensity

Quadrant light percentage fusion

Current Voltage

Real-time bus power monitoring

Active Session Logs

Real-time green-coded event stream tracking hardware commands & sensor deltas.

Geolocation Lat/Lon Lock

Global coordinate registry

Cloud & Precipitation Sync

Weather hourly logic fetch

Static Gyro Offset Baseline

MPU calibration profile

MPU vectors per hour sweep

Sun Path Analysis

NTP High-Precision Sync

Time & Date sync

Anomaly Detection Reference

Humid & Temp monitoring

Logic engine status check

AI Analysis engine

WiFi SSID

Network connectivity target

Sunrise & Sunset

Daylight boundary timing

Dual Bridge Connectivity

The Hub utilizes a prioritized dual-path communication strategy to maintain hardware links:

Method 2 (Registry LAN): The primary bridge. Searches for the device's registered IP in database.db on the local network.
Method 1 (Direct AP): The emergency fallback. If the registry IP is unreachable, the system prompts a bridge via the Tracker's Direct WiFi Hotspot (192.168.4.1).

Discovery & Registry

Controller Configuration

Local Config Loading

The system reads config.json from SPIFFS (ESP32) or local storage (Pi). Key parameters include Geolocation, Timezone, WiFi SSID/Password, and MPU orientation offsets.

Motion Calibration Table

Define the 9-point reference grid for the specific panel installation:

ID	Compass Point	X (Azimuth)	Y (Elevation)
1	Center	0	0
2	North	0	45
3	North East	32	32
4	East	45	0
5	South East	32	-32
6	South	0	-45
7	South West	-32	-32
8	West	-45	0
9	North West	-32	32

IP Registration

The hub retrieves its dynamic IP and registers with the Central Registry. This allows the Dashboard to tunnel into the local control WebSocket.

Real-time Logic

Online Mode Operations

Predictive Fetching

The hub pulls data at intervals defined in the local config (e.g., every 60m):

Dynamic Intervals: Hub fetches frequency and API endpoint links from config.json.
Payloads: Weather Forecast, Sun Azimuth, and Elevation.

Weather Forecast Sun Azimuth Sun Elevation

Telemetry Persistence & Accuracy

X/Y Angles, Voltages, and LDR deltas are logged to local storage. Syncing & Mirroring ensures data safety: since offline trackers delete old logs when storage is full, the Hub proactively mirrors telemetry to guarantee a permanent system audit trail.

Current Angle SPA Process

Utilizes the Unified SPA API for high-precision solar positioning. In Online Mode, the MPU assist calculates currently sent vectors, creating hourly target angles between sunrise and sunset. Unlike Offline Mode which uses "last known" data (yesterday), Online SPA is refreshed every date to ensure peak seasonal alignment.

Manual Angle Override

Allows operators to bypass autonomous tracking and Manual Set the Current Angle (X and Y) at will via the dashboard. This converts the system into a remote-controlled panel for specific orientation tasks or emergency positioning.

MPU Assist (Pattern Engine)

The MPU Assist utilizes the current geolocated coordinates to calculate sunrise and sunset times. It generates a series of hourly MPU target angles derived from the Basic SPA model. In Online Mode, this calculation is performed fresh daily, ensuring the tracker has the latest seasonal vector path without relying on historical "learned" data.

Hub Intelligence & Safety Alerts

Exclusive to the ANN Variant in Online Mode, this autonomous system is integrated with Fail-Safe protocols to monitor mechanical integrity. While the basic Voltage Watchdog is universal to all hardware, only ANN Hubs trigger these smart alerts:

Mechanical Stuck: Motor active but no MPU change detected.
Target Not Reached: Alignment failure after max adjustment cycles.
Emergency Dispatch: Hardware overrides during critical failures.

Online Feature Scaling

A breakdown of how each variant performs when connected to the SkyLink API:

Variant	Online Enhancement	Operational Logic
FIXED	MPU Assist	API retrieves coordinates -> Checks Database Pattern lookup -> Generates recalibration angles based on live Sun Rise/Set.
CONVENTIONAL	LDR Follow Mode	Initially uses MPU Assists -> LDR Follow Mode Engine -> Logs successful vectors to ANN Dataset for future pattern learning.
ANN	AI Analysis Engine	Predictions based on Weather API + Accuracy logs -> Adjusts MPU vectors even when cloudy to maintain peak hypothetical output.

Hub Intelligence Alerts

Autonomous notification system triggered by Weather AI Analysis and mechanical state checks:

Mechanical Stuck: Triggered if the motor is active but a change in MPU angle is not detected.
Target Not Reached: Fired if the panel fails to align within the error margin after maximum adjustment cycles.
Weather Warnings: Notifications based on Rain/Storm predictions from the online suite.

Autonomous Operation

Offline Operation Modes

Current Angle SPA Process

Standard Basic SPA Cycle calculations performed locally using geocoordinates. In this mode, the system relies on Last Known Good data (e.g., yesterday's sun path) if a new daily update cannot be fetched via API.

LDR Follow Mode Engine

This engine operates independently of the API, allowing the controller to track light using local sensors.
Logic Path: Starts from Predicted Sunrise -> Scans for best LDR Delta (Min 90% threshold) -> Continuously tracks until Sunset Edge.

Normal Mode: Maintain >90% detected light intensity relative to the predicted SPA vector.
Cloudy Fallback: If light falls <90%, orient to the highest detected light specifically near the current drift angle to stay on the Sun's path.
Aggressive Mode: A high-sensitivity state where the controller follows the absolute brightest light source detected, regardless of the current predicted solar angle.

ANN Offline Dataset Prediction

For ANN variants, the system performs "Incremental Learning" by eating previous datasets. It utilizes high-accuracy tracking data logged during **Conventional Mode** to generate future weight predictions even while offline.

ANN DATASET FETCHED FROM CONVENTIONAL

Hour	LDR Diff	MPU X/Y	Acc %	Az/El
9:00	23	35.2 / 20.1	97%	110 / 34
10:00	2	36.0 / 22.5	88%	120 / 42
11:00	1	37.1 / 23.0	85%	130 / 50

Telemetry & Accuracy Tracking

Focuses on Local Buffering and Accuracy Scoring. Logs stored in circular buffers are processed against internal LDR peak intensities to determine performance metrics.

Self-Calibration & Safety

The core Fail-Safe & Watchdog system. Features RTC Cycling, Safe Tilt Cycles for logical conflicts, and hardware stall protection.

Fail-Safe

Connectivity Fallback

Connectivity Watchdog

If the API heartbeat is lost for >10 mins, the hub reverts to Fallback Mode.

Time-of-Day Cycling

Uses internal RTC to estimate sun position based on the seasonal table:

Morning: Positive Y (East focus)
Midday: Y ≈ 0 (Zenith focus)
Evening: Negative Y (West focus)

Architecture Delta

Variant Comparison Matrix

Category	FIXED	FIXED ONLINE	CONVENTIONAL	CONV. ONLINE	ANN	ANN ONLINE
Current Angle SPA Process	SPA	SPA	SPA + LDR Follow Mode	SPA + LDR Follow Mode + MPU Assist	ANN Integrated SPA + LDR Follow Mode	ANN Integrated SPA + MPU Assist + LDR Follow Mode
LDR Follow Mode Engine	Disabled	Disabled	Enabled	Enabled	Enabled	Enabled
Manual Override	Disabled	Enabled	Disabled	Enabled	Enabled	Enabled
Telemetry & Accuracy Tracking	Local Buffer	Logs API Mirroring	Accuracy Scoring	Logs API Mirroring	Dataset Persistence	Logs API Mirroring
Fail-Safe & Watchdog	Voltage Watchdog	Voltage Watchdog	Voltage Watchdog	Voltage Watchdog	Voltage Watchdog	Voltage Watchdog + Safety Alerts