Building an IoT-Based Smart Conveyor for Quality & Inventory Control
How we engineered an ESP32 and Arduino-driven smart conveyor prototype combining real-time load cell calibration, ultrasonic counting, robotic sorting, and cloud logging.
IoT-Based Smart Conveyor System for Integrated Quality, Inventory and Process Control
Md. Muqtadir Fuad, Abdullah Al Mazid, Md. Ahabab Ul Haq, Most. Noshin Ibnat, Chowdhury Mouinur Rahman & Nafis Ahmad
In labor-intensive manufacturing facilities, manual weight checking and piece counting are notorious sources of bottlenecking and variability. Operators experience fatigue over extended shifts, causing out-of-spec products to pass inspection and creating discrepancies between actual production output and logged inventory.
To address this operational challenge, we designed and fabricated an IoT-Based Smart Conveyor System. The prototype demonstrates a cost-effective, automated solution for real-time weight discrimination, automated item counting, mechanical rejection/sorting, and cloud-synced inventory telemetry.
Electromechanical Hardware Stack
The system integrates sensing, actuation, and computation into a cohesive feedback loop:
- Weight Measurement: A 5kg strain-gauge load cell connected to an HX711 24-bit precision analog-to-digital converter (ADC) for high-resolution weight measurements.
- Object Detection & Counting: An HC-SR04 ultrasonic sensor positioned above the belt to detect item passage and trigger visual verification indicators.
- Actuation & Sorting: An SG-90 micro servo motor actuating a robotic gate arm to route defective or out-of-spec packages onto a bypass lane.
- Conveyor Drive: A 6V 600 RPM high-torque DC gear motor modulated by an L298N dual H-bridge motor driver for smooth acceleration and stopping.
- Dual Microcontroller Architecture: An Arduino Uno handling real-time low-level sensor sampling and motor PWM, coupled with an ESP32 for Wi-Fi telemetry and HTTP data streaming.
Control Logic & Operational Sequence
Weight calibration is performed by converting raw 24-bit counts from the HX711 ADC into physical mass (grams) using a linear tare-offset and scaling factor:
Load Cell Mass Calibration Formula:
Rejection Condition:
The system operates under an event-driven control sequence to ensure minimal idle power consumption and reliable sorting:
- Idle State: The conveyor remains stationary until an object is placed on the infeed weighing platform.
- Weight Acquisition & Thresholding: The HX711 samples the strain gauge. If the measured mass exceeds the operational trigger threshold ($W \ge 200\text{ g}$), the system evaluates tolerance limits.
- Conveyor Engagement: The L298N driver engages the DC motor, advancing the item down the track.
- Passage Verification: The ultrasonic sensor detects object transit, increments the validated batch counter, and pulses an on-board indicator LED.
- Rejection / Routing: If the reject flag is triggered, the ESP32 signals the SG-90 servo to deflect the item before it reaches the final bin.
- Telemetry Broadcast: The ESP32 formats the payload (timestamp, piece weight $W$, pass/fail flag, total tally) into a JSON packet and transmits it to the centralized web server.
Video Demonstration
Performance & Key Takeaways
During experimental trials, the automated conveyor achieved a 100% item count accuracy and consistent weight-sorting responsiveness under high cycle rates. The direct integration of IoT telemetry eliminated paper-based logging delays, demonstrating how accessible embedded electronics can upgrade conventional manufacturing lines to Industry 4.0 standards.
References & Citation
For full technical specifications, circuit schematics, and experimental data, consult the published technical poster:
- DOI: 10.13140/RG.2.2.25476.23683
- ResearchGate: View Publication Archive
- Video Demonstration: Watch on YouTube