Accessing Physical Document through IoT-Enabled Smart Tagging

Accessing Physical Document through IoT-Enabled Smart Tagging

Shalini Bhartiya

Vivekananda Institute of Professional Studies-Technical Campus, New Delhi, India GuruGobind Singh Indraprashtha University, New Delhi, India

Abstract: Despite rapid digitalization, the storage of documents and files in physical form continues to remain prevalent. With time, the shelves, racks, cupboards or any other medium of storage become cluttered and difficult to manage. Hence, automated tagging of physical documents offers greater flexibility, efficiency, and adaptability compared to traditional manual methods of organizing and searching records. This paper identifies available techniques for tagging physical documents and proposes an Internet of Things (IoT)enabled automated system for real-time document tracking. By utilizing low-cost hardware components and a scalable system architecture, the proposed model facilitates quick and direct file retrieval, significantly reducing manual effort and search time.

Keywords: Automated Tagging, Internet of Things (IoT), Physical document, Physical Storage, Tracking

1. INTRODUCTION

   Physical files such as printed reports, receipts, and logbooks are critical, as they serve as official evidence in audits, legal proceedings, and quality assurance processes. Domains like healthcare, banking, Financial service providers, postal services, government departments, law firms, courts, educational institutions, and many other preserve the structured/unstructured data in physical form. The government and various official bodies mandates every organization to keep physical copies of their records for a stipulated period of time. Almost in every organization, the data lies in physical forms for various audit purposes and references. In general, how the data is organized physically? Usual trend is pinning the documents in files with descriptive labelling on the file covers. But, the question is, Is this method efficient while accessing the relevant data without hassle? Manual file tracking and retrieval processes are not only time- consuming but also prone to errors, leading to inefficiencies, misplaced files, and operational delays1,2. Large Scale organization are capable to invest and manage such data using cost extensive but efficient techniques and technologies. But the system is compromised in Medium and Small Scale organizations that struggles with the funds and competitions to sustain. Responsible personnel are tasked with verifying document completeness, adherence to standardized templates, and proper categorization using color-coded labels, digital labels or

   self-devised categorization.

   The management and retrieval of physical documents3 in storage facilities present a persistent challenge, especially in dynamic and high-volume environments. Locating physical documents remains a time-consuming

   task, particularly in large archives. Frequent delays and inefficiencies in retrieving files, especially in the absence of a systematic real-time tracking mechanism is the common scenario in almost every organization.

   Recent advances in the Internet of Things (IoT), Artificial Intelligence (AI), and edge-cloud computing are transforming this landscape. Modern approaches aim to build integrated ecosystems where sensor networks, RFID tags, Bluetooth Low Energy (BLE) beacons, and edge devices collectively enable real-time file localization, status tracking, and predictive alerts4-6. These architectures are further enhanced by AI models capable of interpreting document metadata, recognizing usage patterns, and supporting intelligent retrieval decisions7-8.

   Despite these innovations, several key limitations persist. Fragmented standards across hardware and communication protocols hinder interoperability9. Energy efficiency remains a major concern for battery- operated and embedded devices10. Additionally, there is a notable lack of scalable, explainable Artificial Intelligence (AI) frameworks that can provide transparent and auditable decision-making for autonomous document handling10-11. As a result, existing systems often lead to confusion, low confidence in the automation, and, in certain cases, redundant development efforts or "reinventing the wheel." To address these issues, a tagging system that facilitate in direct access to the required file or document is proposed. This study introduces a low cost Internet of Things (IoT)-based solution using NodeMCU ESP8266 modules arranged in a mesh network. This network forms a distributed and fault-tolerant communication infrastructure that connects racks of files to a centralized

   control system.

   The paper is divided into six section. Section 1 lays the

   foundation of the work further studied for its relevance identifying the need and challenges from the previous research works in Section 2. Section 3 describes the hardware and software utilized to devise a probable solution for automated tagging of physical documents. Section 4 proposes the model designed and illustrates the functional implementation of the model whose viability and significance is discussed in Section 5. Section 6 concludes the work providing the key benefits of the proposed solution and identifying future directions of research.

2. LITERATURE REVIEW

   The literature study assembles on automating the identification, organization, and retrieval of physical or unstructured digital documents by interweaving IoT connectivity, lightweight micro-controllers, and AI-driven analytics. Early work on Radio Frequency Identification (RFID) and barcode system12-14 demonstrated tangible efficiency gains in libraries, hospitals, and warehouses but were constrained by line-of-sight reading and costly hardware. Subsequent studies added wireless intelligence: Use of Message Queuing Telemetry Transport (MQTT) for real-time home communication by15, mapping the architectural and protocol foundations9,16-17 enabled later IoT projects to exploit further. Bluetooth Low Energy (BLE) beacons18 and hybrid BLE-RFID designs6 extended range and context awareness, yet still lacked visual localization cues.

   Node-level innovation accelerated after 2020. Transformers19 and ESP-based mesh networks10,20 for resilient document tracking were identified while paired Wi-Fi or LED indicators17,21-22 with web dashboards, to close the humanmachine gap. Mobile-friendly locators built by23-24 cut retrieval time but relied on centralized clouds, raising latency and scalability issues. Comparative hardware evaluations25 confirmed NodeMCUs cost advantage, and how mobile robots can extend sensor26 reach in smart archives.

   Concurrent research has focused on content understanding. AI methods used for unstructured document analysis, whereas multiple survey27-29 highlighted how large language models reshape document automation pipelines. Practical NLP-based product tagging and compared tag- versus folder-based30 knowledge organization illustrates underscoring usability trade-offs.

   Across logistics31 synthesized automation trends, noting fragmented deployments and piloted IoT library automation that remains limited by the absence of mesh or LED-based cues. Cutting-edge sensor designs, such as the self-powered paper security tag push sustainability and interactivity forward but still face range and durability hurdles. Secure indexing32 and government-grade RFID tracking is addressed yet without integrating user-friendly localization aids.

   The area under study is steadily moving from isolated

   RFID/Barcode solutions toward integrated, mesh- networked, AI-enhanced platforms tha combine real-time visual cues, mobile interfaces, and cloudedge intelligence28 to make physical-document retrieval as seamless as digital search.

   There remains a lack of fully integrated solutions combining mesh networking, Graphical User Interface (GUI) interfaces, real-time LED localization, and decentralized communicationall of which are addressed in this proposed model.

   1. Comparison Table: Prior Studies vs. Proposed System

      An extensive study of peer-reviewed research papers gave an overview of technologies used for the topic under study. A comparison matrix evaluates the existing system based on key performance indicators such as retrieval speed, accuracy, scalability, ease of integration, power efficiency, and AI integration. Also, the matrix reflects the key benefits of the proposed model in comparison of the existing systems. This strengthens the belief and confidence to carry forward the idea and design the proposed model.

      Feature/Criter ia	Prior Studies (RFID/BLE/Barcode/IoT )	Proposed System
      Real-time Visual Localization	Limited or absent	LED indicators with NodeMCU
      Mesh Networking	Rarely implemented	Wi-Fi based NodeMCU mesh network
      Editable GUI	Basic or non-existent	Fully editable and user- friendly GUI
      Cost-effective Microcontroller s	Some used expensive platforms	NodeMCU ESP8266 (low-cost)
      Real-time Feedback Loop	Often delayed or passive	Immediate LED signal upon search
      Decentralized Architecture	Mostly centralized	Distributed mesh with fault tolerance
      Scalability Tested	Limited rack simulation	Scalable across multiple racks/shelves
      Energy Optimization	Rarely addressed	Future scope included
      Modular and Maintainable Design	Often static	Modular IoT nodes on racks
      Performance Metrics &	Seldom presented	Includes detailed latency mathematical

      Table 1. Comparative Study of Proposed Model with Prior Studies17-18, 30-35

      Feature/Criter ia	Prior Studies (RFID/BLE/Barcode/IoT )	Proposed System
      Latency Modelling		model

3. METHODOLOGY

   Based on identified gaps, a unified framework is designed that supports the real-time location tracking of the documents. The IoT infrastructure includes: NodeMCU8266, LED, Register, Battery(Rechargeable), Battery Case, Switch and PCB Board. The LEDs can be increased/decreased as per the requirements.

   NodeMCU is based on the ESP8266 chip33, which has native Wi-Fi support. It eliminates the need for external Wi-Fi modules like ESP01 or Ethernet shields. It enables low-cost deployment of multiple nodes in mesh networks or sensor grids. It is suitable in document tracking and is enough for small- to medium-scale control systems like LED-based document indexing. With the support of ESP- NOW mesh control it can efficiently be used in distributed systems for applications like decentralized file search or environmental monitoring.

   A mesh Network using ESP-NOW Mesh network protocol to be setup enabling centralized control over distributed LED nodes. One NodeMCU (Gateway) is set as the central controller (Figure 1). This is a super-fast, direct way for NodeMCUs to send small messages to each other without needing a Wi-Fi router. It will provide a robust system for wirelessly managing multiple LED- equipped NodeMCU devices.

   Each ESP8266-based NodeMCU (Remote) is physically installed near or inside the storage units. Upon receiving a command from the central system (through the Python- based GUI (Figure 1)), the NodeMCU activates corresponding LED to visually guide the user to the document's exact location. This immediate feedback significantly reduces search time, especially in environments with a high volume of physical files.

   This setup allows the GUI to discover active NodeMCUs, send specific LED commands to any target node, and receive real-time status updates.

   1. System Overview

      Let:

      • G: Gateway NodeMCU, connected via USB to PC

      • R = {R, R, …, R}: Set of Remote NodeMCUs, powered independently

        Fig 1: IoT-based Hardware Architecture of Smart-File Tagging

        Each node (gateway or remote) controls:

        • L = {L, …, Ln}: LEDs via GPIO pins D = {D, …,

          Dn}

        • S(t) = {s(t), …, sn(t)}: LED state vector of node

          k at time t

          – k {G, 1, 2, …, n}

   2. Command Model

      Command issued from the Python GUI to the Gateway:

      C(t)=(Nk,i,a)

      Where:

      • Nk: Target node ID (Gateway or Remote)

      • i{1,…,n}: LED index

      • a{0,1}: Action (ON = 1, OFF = 0)

      The gateway sends N via WiFi using ESP-NOW or mesh protocol. Remote Node parses and updates:

      Sij(t+1)=Command received from G

   3. Node Decision Logic

      Each node N evaluates:

      (C(t)) = s(t+1) = a if N = k

      Ignore otherwise

      Gateway Node G:

      • If N = G, it executes LED state change locally.

      • Else, it wirelessly forwards the command to the corresponding remote node R.

   4. Wireless Command Propagation

      C(t) where N G, G R (via WiFi Mesh)

      Remote node R applies:

      s(t+1) = a upon receiving C(t)

   5. State Transition Function

      The global LED state at time t is:

      S_global(t) = {S(t), S¹(t), …, S(t)}

      Given C(t), the updated state is:

      S_global(t+1) = T(S_global(t), C(t))

      Where T is the composite state transition function for all nodes.

      , = ()

      4 PROPOSED MODEL

      The proposed model presents an intelligent system for rapid file retrieval using IoT-based smart document tagging and tracking system. It integrates hardware components, communication protocols, and a user interface to form a cohesive and intelligent document management solution. The microcontrollers are organized into a mesh network, enabling peer-to-peer

      = {

      (1)

      communication and scalability across multiple nodes.

      Once the controlling node is identified, it activates the associated LED indicator, denoted as:

      _ = _

   6. System Latency Model

      This activation helps visually locate the correct rack. The total time taken to perform this operation is defined as:

      Ttotal=Tdb+Tnet+Tled

      Where:

      • Tdb: time to retrieve (fj) from database

      • Tnet: time to transmit signal si over the mesh network

      • Tled: time to activate LED

   7. Optimization Goal

      min(Ttotal)

      This represents the optimization goal of minimizing the total time Ttotal, which includes time from database lookup, network communication, and LED activation.

   8. Scalability & Reliability Constraint

Average latency per node should be within th allowed threshold:

1. Architecture of the Proposed Model

   The system is structured around a NodeMCU8266 microcontroller installed as Gateway Server, chosen for its low cost, Wi-Fi capabilities, and ease of integration. Each storage location (e.g., cupboard section, shelf, racks) is equipped with a LED indicator connected to Remote NodeMCU8266(Figure 2).

   () L(n): Load (or tasks/connections) on node n Lmax: Maximum allowable load per node

   N: Set of all nodes

   n N: This condition applies to every node in the network

   The system remains scalable as the number of nodes increases, where the load or number of connections each node can handle might be restricted.

   Signal propagation probability must remain above a minimum reliability threshold:

   ()

   P(n): The signal propagation probability from node n. It represents how likely it is that a message or signal sent by node n successfully reaches its intended destination (such as another node or central controller).

   Pmin: A minimum acceptable threshold for signal reliability depending on system requirements.

   Fig 2: Proposed Model for Smart Tagging of Physical

   Documents

   1. Gateway NodeMCU:

      NodeMCU ESP8266 Board

      USB Port: Connect a Micro-USB Cable from the NodeMCU's USB port to an available USB Port on the computer. (This provides power to the NodeMCU and establishes the serial communication link for the Python GUI.)

   2. Remote NodeMCU(s) (Wireless Mesh Nodes):

      • NodeMCU ESP8266 Board

        • Power Supply: Connect a 5V Micro-USB Power Adapter (like a phone charger) to the NodeMCU's USB port. (These nodes are powered independently and communicate wirelessly.)

      • LED Connections: (These are the LEDs that will be controlled by commands received wirelessly from the gateway NodeMCU.)

        LEDs (x6):

        • § Each LED's Long Leg (Anode) connects

        • to a 220 Ohm Resistor.

        • § The other end of the 220 Ohm Resistor connects to one of the NodeMCU's Digital Pins(D1. Dn).

        • § Each LED's Short Leg (Cathode) connects

        directly to a GND (Ground) Pin on the NodeMCU.

   3. Racks (or any other medium of Physically placed files/documents)

      • Each Rack is fitted with Remote NodeMCU

        ESP8266

      • Wired LED indicators are fitted on each shelf for recognizing the relevant area of search.

   4. A Python-based Graphical User Interface (GUI)

      Allows the user create the database of the documents/files/books etc. Each document is tagged with a unique ID and associated rack and metadata including name, date, category etc. The user can edit existing records and delete outdated or irrelevant/redundant entries. The server once configured, would communicate with the master NodeMCU (Gateway) using Hyper Text Transfer Protocol(HTTP) or MQTT protocols [6] to transmit control commands.

      The GUI serves as the central control hub for the proposed IoT-enabled document tagging and tracking system. It leverages lightweight, open-source libraries such as Tkinter (standard GUI package) or PyQt (for more advanced UIs), providing a user-friendly and platform-independent interface to facilitate real-time file management and interaction with the IoT hardware.

2. Working of the Proposed Model

The serial Gateway connection manages the bidirectional data flow between the Python GUI on the PC and the designated gateway NodeMCU via a serial (USB) connection, handling connection, disconnection, and data encoding/decoding. ESP-NOW Mesh messaging enables direct, fast, and low-power wireless communication between all NodeMCUs in the network, facilitating the exchange of heartbeats and LED control commands without a central router. Node discovery & presence tracking allows each NodeMCU to broadcast its presence (heartbeat) and for the gateway NodeMCU to maintain a dynamic, up-to-date list of all active nodes in the mesh, including their unique identifiers (Node ID, MAC address(Figure 3)).

The active node periodically sends the current list of discovered active mesh nodes from the gateway NodeMCU to the Python GUI, enabling the user to see

which devices are online. Command Forwarding & Execution receives LED

control commands from the Python GUI, forwards them via ESP-NOW to the intended target NodeMCU in the mesh, and allows the target NodeMCU to execute the command by activating the specified LED. Non-Blocking LED control manages the temporary activation of LEDs on NodeMCUs using millis()-based timers, ensuring that the device remains responsive to other network messages and avoids automatic resets.

The system maps each file fj to a specific rack ri, controlled by a node ni, via the mapping function (fj) (Eq. 1). Once the controlling node is identified, it activates the associated LED indicator LEDi as configured. Once active,

i.e. the LED blinks, it helps to visually locate the required file from among the placed racks.

The total retrieval time Ttotal is minimized under the scalability and reliability constraints. With the maximum number of nodes to Lmax such as latency L(n) Lmax and transmission reliability P(n) Pmin, the model is believed to remain scalable and optimized.

Fig 3: Python-based GUI for real-time file retrieval

5 DISCUSSIONS

An IoT-driven model designed to automate the process of tagging and searching the physical documents in real time demonstrates a novel integration of database querying, wireless mesh communication, and localized LED signalling. The data query is made through Python- based GUI. The prototype claim reduction in search time as compared with manual search, especially where the data is stacked and pile of records lie in unstandardized manner. The use of low-power ESP8266 NodeMCU modules in a mesh network allows decentralized communication, ensuring that signals are efficiently routed even in environments with limited infrastructure. It thereby brings significant advantages over traditional point-to-point or centralized architectures.

The model satisfies the key criteria -efficiency,

scalability, and robustness, in automating document tracking in environments with high document volume or physical clutter. With quick response time Ttotal and controlled network performance constraints, the architecture supports real-time, reliable responses with minimal delays. The use of NodeMCU ESP8266 devices are cost-effective, Wi-Fi capable and ease to deployment in mesh configurations. It also supports real-time activation of LEDs associated with the file's physical location, improving human-machine interaction and minimizing search effort. This makes the model particularly effective in libraries, warehouses, hospitals, legal archives, educational institutions or any such organization where physical documents are still prevalent and rapid access is essential.

6 CONCLUSION AND FUTURE SCOPE

In-spite of structured labelling of files, locating them from the store house has always been a tedious and cumbersome task. The proposed model offers a scalable, low-cost, and efficient solution to the persistent problem in large or disorganized storage systems. Combination of database-driven logic, mesh communication, and visual cues has benefitted in enhancement of operational efficiency and user experience. The mathematical foundation ensures that performance remains optimal within defined thresholds, balancing speed, scalability, and reliability.

However, the implementation also highlights several critical insights. hile the model is functional in a controlled setting, secure communication, node-level errors, backup or parallel setup in case of a failure of controlling node, are few practical concerns that requires further deliberations. Also, the system is limited to implement based on signal propagation in a confined physical location. The need for real-time response places a demand on the system's ability to handle multiple concurrent queries, which can be addressed through future integration of edge computing or lightweight task distribution algorithms.

In essence, the model contributes a modular and extensible architecture that can be deployed with minimalistic effort and complexities. It sets a pathway toward developing a hybrid ecosystem where IoT, embedded systems, and information retrieval coalesce to support faster, more accurate, and user-friendly interactions with physical storage environments. Exploring and incorporating technologies like AI, Computer vision, Robotics would provide much intelligent solutions but also compliment with higher development cost, skills and maintenance. Though, with rising demand and acceptance, these models can be evolved in further research work.

ACKNOWLEDGEMENTS

Author SB expresses gratitude to the Vivekananda Institute of Professional Studies-Technical Campus for

providing platform to conduct the research work and to Vital Informer, Monthly Medical Newsletter, New Delhi, India, for the financial support and site for demonstrating the prototype under the grant number VI-MN-2025-RTD- 009.

Conflict of Interest: No conflict of interest existed, as declared by the author.

REFERENCES

1. Sharma, N., & Singh, A. (2018). IoT-enabled file tracking for government offices using RFID. International Journal of Internet of Things, 7(2), 4550.

2. Wamba, S. F., & Chatfield, A. T. (2009). A contingency model for creating value from RFID supply chain network projects. Journal of Strategic Information Systems, 18(1), 3044.

3. Hartzler, A., Jones, W., Klasnja, P., & Bruce, H. (2009). Better to organize personal information by folders or by tags?: The devil is in the details. Proceedings of the American Society for Information Science and Technology, 45, 113. https://doi.org/10.1002/meet.2008.1450450214

4. Bayani, M., Segura, A., Alvarado, M., & Loaiza, M. (2018). IoT- based library automation and monitoring system: Developing an implementation framework. E-Information Sciences, 8(1), 83100.

5. Ferreira, G., Opinião, A., Das, S., Goswami, S., Pereira, L., Nandy, S., Martins, R., & Fortunato, E. (2022). Smart IoT enabled interactive self powered security tag designed with functionalized paper. Nano Energy, 95, 107021.

   https://doi.org/10.1016/j.nanoen.2022.107021

6. Rajesh, S., Menon, V., & Iyer, N. (2021). BLE and RFID based hybrid tracking system for educational records. International Journal of Computer Applications, 183(45), 3439.

7. Liu, C.-L., Hsaio, W.-H., Lee, C.-H., & Chen, C.-H. (2013).

   Clustering tagged documents with labeled and unlabeled documents. Information Processing & Management, 49(3), 596 606. https://doi.org/10.1016/j.ipm.2012.12.004

8. Mahadevkar, S. V., Patil, S., Kotecha, K., et al. (2024). Exploring AI-driven approaches for unstructured document analysis and future horizons. Journal of Big Data, 11, 92.

   https://doi.org/10.1186/s40537-024-00948-z

9. Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., & Ayyash, M. (2015). Internet of Things: A survey on enabling technologies, protocols, and applications. IEEE Communications Surveys & Tutorials, 17(4), 23472376.

10. Tran, H. Q., Le, D. M., & Nguyen, V. H. (2022). Energy-efficient wireless mesh tracking for dense file archives. Journal of Embedded IoT Systems, 8(3), 123132.

11. Singh, T., & Batra, H. (2023). AI-assisted RFID systems for legal document access. Journal of Artificial Intelligence in Law, 4(2), 8997.

12. Karthikeyan, V., & Gnanasekar, R. (2014). Design of barcode- based patient record tracking system. International Journal of Biomedical Engineering and Technology, 14(1), 5765.

13. Kumar, R., Goyal, P., & Bansal, N. (2012). RFID based library management system. International Journal of Advanced Research in Computer Science and Software Engineering, 2(8), 9195.

14. Zhang, L., Wang, H., & Xu, F. (2020). Cloud-assisted RFID

    systems for library inventory and tracking. Library Technology Reports, 56(3), 1726.

15. Piyare, R., & Lee, S. R. (2013). Real-time communication system based on MQTT protocol for smart home. International Journal of Smart Home, 7(1), 161168.

16. Wang, Q., Alyahya, S., Bennett, N., & Dhakal, H. (2015). An RFID-enabled automated warehousing system. International Journal of Materials, Mechanics and Manufacturing, 3(4), 287 291.

17. Want, R. (2006). An introduction to RFID technology. IEEE Pervasive Computing, 5(1), 2533.

18. Roy, A., Narayan, S., & Mehta, P. (2017). Smart warehouse asset locator using BLE beacons. IEEE Sensors Journal, 17(15), 4881 4888.

19. H. Jang, S. Choi, E. Kwon, & C. Kwon. (2020). IoT device auto- tagging using transformers. In 2020 12th International Conference on Advanced Infocomm Technology (ICAIT) (pp. 4750). IEEE. https://doi.org/10.1109/ICAIT51223.2020.9315384

20. Khan, A., Baig, S., & Javed, T. (2022). ESP-based mesh IoT tracking system for document storage management. IEEE Access, 10, 1578815796.

21. Li, Y., & Wang, J. (2023). Web-integrated Wi-Fi alert systems for physical file monitoring. Smart Infrastructure Systems, 11(4), 215 225.

22. Obeng, A., Kusi, P., & Boateng, S. (2023). Distributed LED notification system using NodeMCU for municipal archives. Journal of Digital Government, 6(1), 8998.

23. Abidin, Z. Z., Zakaria, N. A., Harum, N., Suhaimi, M. S., Abal Abas, Z., Idris, A., Wan Mohd Ghazali, K., & Abdul Hamid, M. H. (2020). Document tracking using Internet of Things devices for fast data retrieval. IOP Conference Series: Materials Science and Engineering, 864(1), 012057. https://doi.org/10.1088/1757- 899X/864/1/012057

24. Ahmed, M., Hossain, R., & Kabir, A. (2024). Real-time document locator using Arduino and Android interface. International Journal of Smart Systems and Technologies, 13(2), 5562.

25. Dey, R., Banerjee, A., & Saha, P. (2020). Comparative evaluation of microcontroller platforms for smart asset tracking. Microelectronics and Embedded Applications Journal, 9(2), 4048.

26. Tuna, G., Gungor, V. C., & Gulez, K. (2013). An autonomous wireless sensor network deployment system using mobile robots for smart environments. Computer Networks, 59, 238252.

27. Achachlouei, M. A., Patil, O., Joshi, T., & Nair, V. N. (2023). Document automation architectures: Updated survey in light of large language models. CoRR, abs/2308.09341. https://doi.org/10.48550/arXiv.2308.09341

28. Oh, K. E. (2024). A comprehensive investigation of researchers shared file management practices in cloud storage. Human Computer Interaction, 120. https://doi.org/10.1080/07370024.2024.231055

29. Abidin Z.Z., Zakaria N.A., Harum N., Suhaimi M.S., Aas Z. A., Idris A., Ghazali K.W.M. and Hamid M.H. A. (2020). Document Tracking using Internet-of-Things Devices for Fast Data Retrieval, IOP Conf. Series: Materials Science and Engineering. 864, 012057. doi:10.1088/1757-899X/864/1/012057

30. Darwin G. Raralio., Wreloyd Q. Manzano., Angelo Nathaniel L. Lomibao. (2025). Development of Document Tracking and Archiving System for Procurement Management Offices Purchase

    Request in Apayao State College, International Journal of Research and Scientific Innovation (IJRSI), 12(7). DOI: https://doi.org/10.51244/IJRSI.2025.120700132

31. Ferreira, B., & Reis, J. (2023). A systematic literature review on the application of automation in logistics. Logistics, 7(4), 80. https://doi.org/10.3390/logistics7040080

32. Geetha, S., Hariharan, R., & Venkatesan, V. P. (2017). Secured indexing and tagging of IoT based device nodes for service based licensing and secured access. In 2017 World Congress on Computing and Communication Technologies (WCCCT) (pp. 6 10). IEEE. https://doi.org/10.1109/WCCCT.2016.12

33. DSouza, R., & Patel, M. (2025). Modular IoT document localization using ESP microcontrollers. Journal of Embedded Systems & IoT, 7(1), 112.

34. Thushara, M.G., Krishnapriya M.S. and Nair S.S. (2017). A model for auto-tagging of research papers based on keyphrase extraction methods, International Conference on Advances in Computing, Communications and Informatics (ICACCI), Udupi, India, 2017,

    pp. 1695-1700, doi: 10.1109/ICACCI.2017.8126087.

35. Mohamed, H.A., Sansonetti G., Micarelli, A.(2022). Tag-Aware Document Representation for Research Paper Recommendation, arXiv. doi.org/10.48550/arXiv.2209.03660