RFID Warehouse Inventory Tracker with Live Stock Dashboard

Small warehouses still count stock by hand — slow, error-prone, and blind between counts. RFID changes the economics: a tag on every bin and a reader at the gate turns material movement into data automatically. This project demonstrates that on a model warehouse aisle: 48 tagged bins, an MFRC522 reader at the gate, and a handheld reader point, all feeding an ESP32. Firmware resolves tag reads to bin identities, tracks check-in/check-out events, and publishes stock levels over Wi-Fi via MQTT. The dashboard shows a live bin stock table with reorder flags, reads-per-hour graphs and unknown-tag alerts. The project is honest about the technology: 13.56 MHz HF RFID reads at centimetres, not metres — the report explains the range physics and why UHF would be needed for dock-door portals, keeping the claims inside what the hardware actually does.

1. RFID tags are attached to the model's bins and registered in the database with SKU, quantity and reorder level.
2. The gate MFRC522 reader continuously polls its field; firmware runs the anti-collision sequence and resolves each tag's UID.
3. On each read, the ESP32 looks up the bin, updates its last-seen timestamp and movement direction, and publishes the event over Wi-Fi via MQTT.
4. The dashboard updates the bin stock table, increments reads-per-hour, and evaluates reorder levels on every event.
5. Presenting a bin at the handheld reader shows its recorded stock on the dashboard's verify panel for cycle counts.
6. Unknown UIDs raise an alert — an untagged or foreign item at the gate — logged with timestamp.
7. The buyer runs the tag registration procedure for all 48 bins, documented with expected read ranges in the manual.

Tech stack:

• ESP32 development board
• MFRC522 RFID reader modules x2
• RFID tags x48 (13.56 MHz)
• Model warehouse aisle with 48 bins
• MQTT broker + web dashboard
• Arduino IDE (C/C++ firmware)
• Tag registration utility
• 16x2 LCD (gate status display)

• [Tagged-bin inventory] Every bin carries an RFID tag bound to its SKU; the database holds bin, SKU, quantity and reorder level.
• [Gate reader point] An MFRC522 reader at the aisle gate logs bins passing in and out with timestamps — movement becomes data.
• [Handheld reader point] A second reader acts as the cycle-count station: present a bin to verify its recorded stock instantly.
• [Live stock dashboard] Bin stock table with reorder flags, reads-per-hour graphs, and unknown-tag alerts, all from live telemetry.
• [Low-stock alerts] When a bin's quantity falls below its reorder level, the dashboard flags it and logs the event.
• [MQTT telemetry] The ESP32 publishes read events and stock updates as JSON over Wi-Fi via MQTT with immediate event delivery.
• [Anti-collision handling] Firmware implements the MFRC522 anti-collision sequence so multiple tags in the field are read in turn, not garbled.

• Working model warehouse aisle (48 tagged bins, 2 RFID readers, ESP32, gate display)
• Complete firmware source (MFRC522 driver, anti-collision, MQTT, inventory logic)
• Live web dashboard (stock table, graphs, alerts) demonstrated with the aisle
• Circuit and wiring documentation
• Tag registration procedure with expected read ranges
• Component list with ratings
• Project report PDF (RFID background, HF vs UHF, anti-collision, inventory systems, methodology)
• PPT presentation for final review
• Viva Q&A preparation document (RFID physics, MFRC522, anti-collision, MQTT)
• Setup and demonstration guide

• 13.56 MHz HF RFID reads at approximately 3-5 cm (datasheet) — bins are presented to readers, not read at dock-door distance; UHF for that is discussed as future scope.
• Metal bins detune HF tags; the model uses non-metallic bins and the manual notes the limitation.
• Tag UID is the identity — cloning or secure authentication is out of scope for the demonstration.
• Stock quantities are updated by configured check-in/out events; the system tracks identity and movement, it does not weigh bins.
• Telemetry needs Wi-Fi; the gate LCD shows last-read status locally without network.