RTOS-Based Edge Monitoring System for Industrial Machines ⚡

This project is an RTOS-based monitoring system for a 3-motor industrial machine e.g. shuttleless weaving loom.

• STM32F4 running FreeRTOS collects motor electrical and temperature data, processes it in real time, detects faults, and logs important events.
• ESP32 gets data via UART, which hosts a local web dashboard to display live machine and motor health information.

The system is non-intrusive and retrofit-friendly, meaning it can be added to existing machines without changing their control logic. It is designed for monitoring and diagnostics, not for controlling the machine (basic ON/OFF is given).

---

⚡ Why This Project?

This is a production-style edge monitoring system, not a simple demo or data logger.

• RTOS-first design Independent tasks for acquisition, processing, faults, logging, and communication ensure deterministic behavior.

• Predictable real-time performance DMA-driven ADC and bounded task execution deliver low jitter under continuous load.

• Robust fault handling Cycle-bounded fault detection with watchdogs, CRC, and non-blocking logging for safe failure handling.

• Scalable architecture Queue-based pipelines allow adding sensors or machines without redesign.

• Edge-focused system design Time-critical logic runs on STM32; ESP32 is used only for visualization.

This project reflects real RTOS design principles used in industrial embedded systems.

---

⚡ System Architecture

The diagram shows task separation, ISR handoff, queues, fault handling, logging, and supervision flow on the STM32.

---

⚡ Tech Stack

• MCU 1 (STM32F446RE, Nucleo)

  • FreeRTOS for task scheduling
  • ADC with DMA for fast sampling
  • Processing tasks for Vdc, Idc, Power, Temp x3
  • UART Tx for inter-MCU comms

• MCU 2 (ESP32)

  • UART Rx + packet parsing
  • Async WebSocket Server
  • Data pushed to Web Dashboard

• Frontend

  • Chart.js for real-time graphs
  • HTML/CSS/JS for clean UI

---

⚡ Fault Detection & Logging

The system detects and logs:

• Over-voltage / Under-voltage
• Over-current
• Over-temperature
• Under-power / abnormal load

Each fault is:

• Detected locally at the edge
• Timestamped
• Logged to flash
• Forwarded to the ESP32 dashboard

Logging is non-blocking and designed to avoid priority inversion.

---

⚡ Performance Characteristics

Metric	Value
Parameters monitored	12 (V/I/P/T × 3 motors)
ADC throughput	>10 kSamples/sec
Fault detection latency	<10 ms
Telemetry update rate	1-5 Hz
End-to-end latency	<100 ms (typical)
Communication path latency	<50 ms
Logged events	>1000
CPU utilization reduction	~30%

---

⚡ Demo

Live Web Dashboard (ESP32):

Real-time visualization of 3-motor machine telemetry (V/I/P/T), fault status, and system health streamed from the STM32 edge node.

RTOS Task Execution Trace (SEGGER SystemView):

FreeRTOS task scheduling, ISR activity, and inter-task communication showing bounded execution time and deterministic behavior under load.