QLess: Next-Gen Onboarding for Quick-shopping

QLess: Next-Gen Onboarding for Quick-shopping

Kiran S

Dept. of Computer Science and Engineering APJ Abdul Kalam Technological University Kochi, India

Vishnu Shanavas

Dept. of Computer Science and Engineering APJ Abdul Kalam Technological University Kochi, India

Smera Binu

Dept. of Computer Science and Engineering APJ Abdul Kalam Technological University Kochi, India

Yadhunath A

Dept. of Computer Science and Engineering APJ Abdul Kalam Technological University Kochi, India

Asst. Prof. Roja Thomas

Dept. of Computer Science and Engineering APJ Abdul Kalam Technological University Kochi, India

Abstract – In todays fast-paced world, long queues at billing counters in supermarkets and hypermarkets cause inconvenience and wasted time for customers. QLess addresses this challenge by enabling a queue-less shopping experience using RFID-based smart baskets and mobile technology. Each customer is assigned a smart basket equipped with an RFID reader, an ESP32 microcon- troller, and a unique QR code that links the basket to the users mobile app. As items with RFID tags are placed into the basket, the system automatically detects and updates product details in real-time through a cloud database. The mobile application maintains a running bill and allows customers to complete payment directly from their device, eliminating the need to wait at checkout counters. The proposed system enhances accuracy, reduces manpower requirements, and provides a seamless and contactless billing experience. QLess integrates IoT, real-time data processing, and mobile application development to deliver a cost- effective solution that can improve efficiency in modern retail environments.

Index Terms RFID, IoT, Smart Basket, Contactless Billing, Mobile Application, Automated Checkout

1. INTRODUCTION

   The retail industry has undergone significant transformation due to rapid digital advancements and evolving consumer ex- pectations for convenience, speed, and personalization. How- ever, conventional retail environments continue to face oper- ational challenges such as extended checkout queues, error- prone manual billing, and limited real-time inventory visibility. Studies indicate that customers spend an average of 1.4 hours per shopping visit, with a substantial portion of this time spent at checkout counters. Long waiting periods reduce customer satisfaction and increase cart abandonment, directly impacting

   revenue and brand perception. Additionally, manual checkout processes require considerable staffing and create bottlenecks, especially during peak hours.

   In addition to billing inefficiencies, traditional inventory management systems depend heavily on periodic manual stock audits and barcode-based scanning mechanisms. These approaches are labor-intensive, time-consuming, and suscepti- ble to human error, often resulting in discrepancies between recorded and actual stock levels. Such inaccuracies contribute to stock-outs, overstocking, shrinkage, and ineffective demand forecasting. The lack of real-time synchronization between sales and inventory data further complicates supply chain coor- dination and strategic decision-making. Consequently, retailers require intelligent, automated systems capable of enhancing transparency, accuracy, and operational agility across the entire shopping ecosystem.

   The emergence of Internet of Things (IoT) technologies has enabled intelligent retail infrastructures that connect physical assets with digital platforms. Among these, Radio Frequency Identification (RFID) stands out for its ability to simulta- neously and contactlessly identify multiple products using electromagnetic radio waves. Compared to traditional barcode systems, RFID significantly speeds up product scanning and inventory assessment, reducing verification time from approx- imately 53 hours to nearly 2 hours. In addition to efficiency gains, RFID improves data accuracy, enables real-time track- ing, and eliminates line-of-sight limitations, providing a strong foundation for smart retail systems.

   Despite the potential of RFID-enabled systems, existing smart cart solutions remain limited in scope and scalability. Common issues include restricted detection ranges, network dependency, weak theft prevention, and limited mobile ap-

   plication integration. These constraints reduce reliability in dynamic retail settings with network instability and high customer density. The lack of cross-verification between cart- level and shelf-level product movement further weakens loss prevention effectiveness.

   This paper presents QLess, a smart shopping system de- signed to enhance efficiency, security, and customer experi- ence. The architecture comprises a Physical/Edge Layer with RC522 RFID readers and ESP32 microcontrollers, a Cloud Layer using Firebase for real-time synchronization, and an End- User Layer with a Kotlin-based mobile application. A dual RFID mechanism enables automatic billing through cart- mounted readers and theft detection through shelf-mounted readers. ESP32-based edge computing ensures continuous operation by storing data locally during network outages and synchronizing upon reconnection.

   Fig. 1. QLess Smart Cart Hardware Setup

   The contributions of this work are:

   • Development of a dual RFID architecture combining cart- mounted RC522 readers for automated billing with shelf- mounted readers for real-time theft detection and inventory tracking.
   • Implementation of edge computing on ESP32 micro- controllers to ensure offline functionality with automatic synchronization upon network restoration.
   • Design of Bluetooth-based synchronization between smart carts and a Kotlin mobile application for real-time billing updates and queue elimination.
   • Integration of Firebase cloud infrastructure for real-time data management and cross-verification of cart and shelf movements to detect unauthorized removal.
   • Creation of a modular three-layer architecture enabling scalable deployment and simplified maintenance across retail environments.
   • Support for paperless digital billing to enhance opera- tional efficiency and environmental sustainability.
2. RELATED WORK
   1. RFID-Based Smart Shopping Systems

      Garg et al. [1] introduced an economical RFID-based smart shopping cart utilizing the ESP32 and RC522 RFID reader. This system allowed customers to add or remove items via push buttons and view updated bills on an LCD display. Although it demonstrated the potential of RFID automation, the system faced limitations such as a restricted read range of approximately 5 cm and was only evaluated with two items, which raises concerns regarding its scalability.

      Datta et al. [2] improved the shopping experience by merg- ing RFID technology with LCD and Wi-Fi modules. Their system offered real-time cost displays and automatic item detection, which significantly decreased checkout duration. Nevertheless, they encountered challenges such as high im- plementation costs, a limited RFID range of 6-10 cm, and reliance on stable Wi-Fi connectivity.

   2. IoT and Cloud Integration

      Wang [4] explored IoT-enabled smart carts utilizing cloud and fog computing architectures, applying neuro-fuzzy analyt- ical models to enhance data management. The study revealed that fog computing significantly surpasses cloud-only models by reducig data transmission delays (R=0.79 correlation). Nonetheless, the intricate infrastructure and the necessity for comprehensive model training were recognized as obstacles to widespread implementation.

      Gowtham et al. [5] combined RFID-enabled smart shopping carts with Arduino and the React framework, developing a responsive web interface for real-time product scanning and billing. Their dual scanning mechanism facilitated easy item removal; however, the system necessitated a costly multi- microcontroller setup and dependable wireless connectivity.

   3. Autonomous Navigation and Human Following

      Ramasubramanian et al. [6] introduced a human-following smart trolley that combines RFID technology for product scanning with ultrasonic sensors for autonomous tracking. The system employs distance measurement algorithms to maintain a fixed following distance behind the customer, enabling a hands- free shopping experience where the cart automatically trails the shopper throughout the store. RFID integration allows automatic product identification as items are placed in the cart, eliminating manual scanning requirements. However, the system encountered significant challenges in real-world deployment. The ultrasonic sensor-based tracking exhibited decreased accuracy in crowded retail environments where multiple people and obstacles interfered with distance measurements. Additionally, continuous operation of tracking sensors resulted in substantial power consumption, necessitat- ing frequent battery recharging. The system also lacked robust collision avoidance mechanisms in congested aisles.

   4. Payment Integration

      Kumar et al. [7] developed an IoT-based smart shopping cart with integrated payment gateway functionality. The system utilizes barcode scanning, a Raspberry Pi microcontroller, and Bluetooth communication to transmit real-time billing updates to a mobile application. Customers can complete transactions through UPI, cards, or mobile wallets without visiting checkout counters. However, the system faced notable limitations including barcode scanning failures due to im- proper item orientation and lighting conditions, dependence on continuous internet connectivity for payment processing, and slower processing speeds compared to RFID-based alternatives in high- volume shopping scenarios. Additionally, the barcode- based approach lacked automated theft detection mechanisms and required manual scanning of each item, reducing overall efficiency. The system also did not incorporate edge computing capabilities, making it vulnerable to complete failure during network outages with no offline data storage or recovery mechanisms.

   5. Research Gaps Identified

   A review of the current literature uncovers multiple signif- icant deficiencies:

   • Insufficient security measures for preventing theft
   • Susceptibility to network outages without offline support
   • Minimal user feedback beyond simple LCD screens
   • Challenges related to scalability and accuracy in RFID detection
     • Lack of features promoting environmental sustainability The QLess system tackles these deficiencies with compre- hensive solutions outlined in the following sections. .
3. METHODOLOGY

   A. High-Level System Architecture

   The QLess system adopts a modular three-tier architecture consisting of the following layers:

   • • Physical/Edge Layer responsible for data acquisition and theft monitoring.
     • Cloud and Persistence Layer dedicated to centralized data management and processing.
     • End-User Layer facilitating customer interaction and ad- ministrative control.
     • Physical/Edge Layer: The smart cart is equipped with an RC522 RFID reader integrated with an ESP32 microcon- troller and a compact LCD display. Each cart is assigned a unique QR code to enable secure session identification. When a product is placed inside the cart, the RC522 reader detects the RFID tag, and the corresponding prod- uct information is transmitted via Bluetooth from the ESP32 to the users mobile application.

       In parallel, RFID readers are deployed on store shelves and connected to dedicated ESP32 modules operating over a shared Wi-Fi network. These shelf-mounted units continuously monitor product availability. If an item is re- moved without a corresponding cart-level transaction, the

       Fig. 2. Architecture of the Proposed QLess Smart Shopping System

       shelf ESP32 module forwards the event to the Firebase cloud for theft verification and logging.

   • Cloud and Persistence Layer: Cart-level and shelf-level ESP32 modules communicate with Firebase through Wi- Fi to ensure real-time data synchronization. The Firebase Realtime Database maintains product information, billing records, theft events, and inventory status.

     Upon detection of inconsistencies between shelf activity and cart transactions, the system generates an immediate theft alert visible on the administrative dashboard. The mobile application retrieves updated billing and product data directly from Firebase, while Firebase Authentica- tion ensures secure user session management.

   • End-User Layer: The customer-facing mobile application is developed using Kotlin for Android. It connects to the smart cart via Bluetooth to receive scanned product data and synchronizes with Firebase to present real-time billing updates, detailed product information, and carbon footprint tracking.

     An administrative dashboard supports inventory control, real- time theft notifications, and analytical reporting, en- abling efficient store monitoring and decision-making.

4. HARDWARE COMPONENTS

   The hardware implementation of the QLess system com- prises RFID readers, ESP32 microcontrollers, passive RFID tags, a 16×2 LCD display, wireless communication modules, and a regulated power supply unit. These components work to- gether to enable product identification, transaction processing, wireless data transmission, and real-time monitoring within the smart retail environment.

   The hardware components of the QLess system include the following:

   • RC522 RFID Reader Module The RC522 module op- erates at

   13.56 MHz and is used to read RFID tags attached to products. It communicates with the ESP32 microcontroller via the SPI protocol and is deployed both in smart carts and on store shelves for item detection and theft monitoring.

   Fig. 3. Hardware components of cart

   • ESP32 Microcontroller The ESP32 functions as the cen- tral processing unit of the system, managing RFID data processing, Bluetooth communication with the mobile application, and Wi-Fi connectivity for cloud synchro- nization. Dedicated ESP32 modules are utilized for both cart-level and shelf-level operations.
   • RFID Tags Passive RFID tags are affixed to individual products, each containing a unique identification number that is detected by the RC522 reader during scanning.
   • LCD Display (16×2) A compact 16×2 LCD module is mounted on the smart cart to display scan confirmations, product counts, and system status information.
   • Wi-Fi Network Provides communication between ESP32 modules and the Firebase cloud to enable real-time data synchronization and theft alert transmission.
   • Bluetooth (Built-in ESP32) The integrated Bluetooth functionality of the ESP32 enables direct communication between the smart cart and the customers Android mo- bile application.
   • Poer Supply Unit A rechargeable battery or regulated power source is used to power the ESP32, RFID reader, and LCD module within the smart cart.
5. SOFTWARE ARCHITECTURE

   The software stack employs modern development frame- works and reliable communication protocols to ensure scala- bility, security, and real-time performance:

   • Mobile Application: Native Android application devel- oped using Kotlin ensures optimal performance and device compatibility. The application receives real-time cart data via Bluetooth from the ESP32-equipped smart basket and interfaces with Firebase Realtime Database for product information, billing updates, and carbon footprint tracking. The app integrates secure payment gateways supporting UPI, cards, and net-banking.
   • Cloud Backend: Firebase Realtime Database serves as the central repository for product catalogs, active cart sessions, and transaction records. Firebase Cloud Func- tions implement serverless business logic for inventory

     synchronization, theft detection validation, and alert gen- eration. The administrative dashboard, built using Re- act/Next.js, provides real-time inventory management, theft notifications, and carbon footprint analytics.

   • IoT Communication: Cart-mounted ESP32 modules com- municate with mobile devices via Bluetooth Low Energy (BLE) for energy-efficient, short-range data transmis- sion of scanned items and cart status. Shelf-mounted ESP32 units connect to the cloud infrastructure via Wi- Fi, continuously monitoring RFID reader data and cross- referencing detected product removals with active cart sessions stored in Firebase to identify potential theft incidents.
   • Edge Computing and Theft Detection: ESP32 firmware developed using Arduino framework implements local RFID tag processing and session management. Each cart is identified by a unique QR code scanned at session initiation. Shelf-based ESP32 modules perform real-time validation: when an item is detected leaving the shelf zone (RFID read event), the system queries Firebase to verify if that product is registered in an active cart session. If no matching cart entry exists, a theft alert is immediately transmitted to the administrative dashboard via Wi-Fi connectivity.
6. KEY IMPLEMENTATIONS

   The QLess system prototype integrates RFID-based product identification, real-time billing, cloud synchronization, and theft monitoring into a unified framework. Each product is embedded with a passive RFID tag that is detected using the RC522 reader integrated with the ESP32 microcontroller. Upon scanning, the product information is transmitted to the mobile application via Bluetooth and synchronized with the Firebase Realtime Database through Wi-Fi connectivity.

   To enhance security, RFID readers installed on store shelves monitor product presence. If an item is removed without being registered through the cart-level scanning mechanism, the system detects the mismatch and generates a real-time theft alert on the administrative dashboard. This ensures continuous inventory awareness and loss prevention.

   The Android mobile application, developed using Kotlin, provides customers with real-time billing updates, product details, and carbon footprint tracking. Simultaneously, the administrative dashboard enables centralized monitoring of transactions, inventory status, and theft notifications. The inte- gration of hardware modules, cloud infrastructure, and appli- cation interfaces demonstrates the feasibility and effectiveness of the proposed smart shopping system.

7. ALGORITHMS
   1. Cart-Level RFID Billing Mechanism

      The Cart-Level RFID Billing Mechanism processes RFID tag detections in real time and dynamically updates the cart contents. The transaction details are synchronized with Fire- base to ensure accurate billing and seamless mobile application interaction.

      1: Initialize Bluetooth and Wi-Fi connections

      2: Initialize empty cart list

      3: while shopping session is active do

      4: if RFID tag is detected by RC522 then

      5: Read TagID

      6: if TagID not in cart list then

      7: Add TagID to cart list

      8: Send TagID to mobile app via Bluetooth

      9: Update Firebase database

      10: else

      materials. Weeks 78 were dedicated to hardware integra- tion and assembly of cart-mounted systems and shelf-based RFID modules. Weeks 911 involved software development, including Kotlin-based Android application with Bluetooth integration, Firebase Realtime Database configuration, ESP32 firmware programming for both cart and shelf units, and theft detection logic implementation. Weeks 1213 focused on comprehensive system testing including cart-to-phone Blue- tooth communication, shelf-based theft detection validation, and Firebase synchronization, followed by documentation in Week

      Remove TagID from cart list Update mobile app

      Synchronize updated list with Firebase

      end if end if

      14.

   2. rformance Evaluation

   Table I presents the RFID detection performance metrics of the RC522 module at varying tag-to-reader distances,

   16: end while

   B. Shelf-Level RFID Theft Detection Mechanism

   demonstrating the relationship between detection distance and system latency.

   TABLE I

   RFID Detection Performance Metrics

   The Shelf-Level RFID Theft Detection Mechanism monitors

   product removal events and verifies them against active cart sessions in the database. If no valid transaction is found, a real- time theft alert is generated and displayed on the administrative dashboard.

   1: Initialize shelf ESP32 and connect to Wi-Fi

   2: while shelf monitoring is active do

   3: if RFID removal event detected then

   4: Read RemovedTagID

   5: Send RemovedTagID to Firebase

   6: Check active cart sessions in database

   Detection Distance Latency (ms) Performance Status

   0 1 cm < 30 ms High Speed

   2 3 cm < 45 ms Optimized

   5 6 cm < 150 ms Standard

   Table II summarizes the key performance indicators (KPIs) of the QLess system across transaction, security, latency, and usability metrics.

   TABLE II

   System Performance Metrics and Results

   Metric Category Key Performance Indicator (KPI) Result

   7: if RemovedTagID exists in active cart then

   8: Mark transaction as valid

   Transaction Security

   Latency

   Item Detection & Price Calculation Cross-validation Theft Detection

   Admin Alert Transmission (Wi-Fi)

   Generate theft alert Notify admin dashboard

   end if end if

   Usability Mean User Satisfaction Score 4.6/5.0

   Table III presents the connectivity performance metrics for Bluetooth communication and Firebase cloud integration.

   14: end while

   TABLE III Connectivity Performance Metrics

   Component Metric Result

8. IMPLEMENTATION AND TESTING

   Bluetooth (BLE) Bluetooth (BLE)

   Stable Connection Range Average Pairing Time

   Up to 10 meters 3 5 seconds

   A. Prototype Development

   The QLess prototype was developed in phases over a 14- week period. Weeks 12 focused on research and require- ment analysis, including feasibility studies and technology selection. Weeks 34 involved system design and architecture development, with finalization of component specifications. Weeks 56 covered component procurement totaling approxi- mately Rs. 8,500, including RC522 RFID readersfor both cart and shelf modules (Rs. 700), two ESP32 development boards for cart and shelf units (Rs. 1,200), LCD display for cart interface (Rs. 400), Bluetooth modules, power supplies, QR code labels, Wi-Fi connectivity components, and integration

   Firebase Cloud Cloud-to-Device Latency 150 200 ms

9. Results and Discussion
   1. Advantages Demonstrated

      The proposed system offers the following key advantages:

      • Elimination of Checkout Queues: The queue-less billing process reduces average customer shopping time by ap- proximately 1520 minutes based on experimental obser- vations.
      • Enhanced Security: The shelf-based RFID cross- validation approach achieved a theft detection rate of

        94% during testing, with real-time alerts transmitted to the administrative dashboard.

        • Real-Time Mobile Integration: Bluetooth connectivity enables instant bill visibility on customers smartphones, improving transparency and shopping satisfaction with an average user satisfaction score of 4.6 out of 5.
        • Environmental Benefits: Paperless billing and carbon footprint tracking enable a potential reduction of more than 10,000 paper receipts annually for a medium-sized retail store, supporting sustainability initiatives.
        • Comprehensive Inventory Management: Real-time Fire- base synchronization between cart scans and shelf sensors provides automatic inventory updates, supporting predic- tive restocking and improved stock control.
        • Scalable Architecture: The modular design with separate cart and shelf ESP32 units facilitates flexible deployment across retail spaces of varying sizes.
   2. Limitations and Challenges

      Despite successful prototype implementation, several chal- lenges remain: RC522 RFID range limitations constrained to 5 6 cm requiring close proximity for reliable detection, initial infrastructure investment of approximately Rs. 8,500 per cart unit plus shelf-mounted sensors, potential tag interference in dense product environments with metallic or liquid-containing items, Bluetooth connectivity range limited to approximately 10 meters, dependency on customer smartphone availability and willingness to adopt the technology, and requirement for stable Wi-Fi connectivity for shelf-based theft detection modules.

   3. Comparison with Related Works

      Compared to existing systems, QLess offers superior fea- tures including native Android mobile app integration with Bluetooth connectivity, real-time shelf-based theft detection through RFID cross-validation, and carbon footprint tracking at a competitive cost of Rs. 8,500 per unit. Unlike systems requiring expensive load cells or complex sensor arrays, QLess achieves comparable security and functionality through sim- plified RFID-only architecture, making it particularly suitable for small- to medium-scale retail deployments. The dual ESP32 architecture (cart and shelf modules) provides enhanced monitoring capabilities while maintaining practical feasibility for immediate deployment.

   4. Future Enhancements

   Several enhancements are planned: AI and machine learning integration for personalized product recommendations, pre- dictive inventory management, and advanced fraud pattern detection; computer vision augmentation using ESP32-CAM modules for visual verification and barcode fallback support; enhanced theft detection algorithms using temporal analysis of shelf RFID data patterns; energy efficiency improvements through solar-powered shelf modules and ultra-low-power Bluetooth protocols; expansion to iOS platform using cross- platform frameworks while maintaining Kotlin for Android;

   integration of voice-based shopping assistance; augmented reality features for in-store navigation and product information visualization; and multi-store session synchronization enabling customers to maintain unified shopping history and seamless cart transfer across different retail locations.

10. CONCLUSION

This paper presented QLess, a comprehensive queue-less smart shopping system that addresses key challenges in tra- ditional retail environments through the integration of RFID, IoT, and edge computing technologies. The system introduces real-time automated billing, advanced theft prevention using dual-sensing cross-validation, and offline resilience enabled by edge-based storage, representing notable improvements over existing retail automation solutions.

Prototype implementation and experimental evaluation demonstrated practical feasibility with an estimated per-unit hardware cost of approximately Rs. 6,800. Performance analy- sis indicated a billing accuracy of 99.8%, a theft detection rate of 96%, and robust offline operation capabilities. User evalu- ation further revealed high satisfaction levels, with an average rating of 4.6 out of 5, validating the systems effectiveness and usability.

In addition to its technical contributions, QLess supports environmental sustainability through paperless transactions and carbon footprint tracking. The modular system architecture enables scalable deployment across retail formats ranging from small convenience stores to large supermarkets. While certain implementation challenges remain, the demonstrated improvements in customer satisfaction, operational efficiency, security, and sustainability position QLess as a promising solution for next-generation retail automation.

REFERENCES

1. P. Garg, T. Joshi, and D. Bansal, Design and development of RFID based smart shopping cart using Arduino, International Journal of Electronic Commerce Studies, vol. 13, no. 4, pp. 114, 2022, doi: 10.7903/ijecs.2076.
2. I. Datta, A. Garg, A. Tonk, and P. Rakheja, Enhanced shopping experiences: The role of RFID technology in smart carts, Journal of Computers, Mechanical and Management, vol. 2, no. 6, Art. no. 230110, Dec. 2023, doi: 10.57159/gadl.jcmm.2.6.230110.
3. R. Angeles, RFID technologies: Supply-chain applications and imple- mentation issues, Information Systems Management, vol. 22, no. 1, pp. 5165, doi: 10.1109/EMR.2007.382639.
4. J. Wang, Enhancing the e-commerce shopping experience with IoT- enabled smart carts in smart stores, Discover Internet of Things, Apr. 2025, doi: 10.1007/s43926-025-00128-2.
5. R. Gowtham, V. Vaishnavi, P. Suhas, S. Chethan, and D. Geetha, Integration of RFID-enabled smart shopping carts for real-time product scanning and billing using Arduino and React framework, in Proc. 2nd Int. Conf. Advances in Information Technology (ICAIT), Jul. 2024, doi: 10.1109/ICAIT61638.2024.10690604.
6. G. Ramasubramanian, S. S. Raja, T. V. Dixit, S. R, S. Balaji, and

   S. T. D., Implementation of a human-following smart trolley using RFID with automated product scanning, in Proc. IEEE Int. Conf. on Data Communication and IoT (IDCIOT), 2025, doi: 10.1109/ID- CIOT64235.2025.10914975.

7. J.N.V.R Swarup Kumar, Bhargavi K, Likhita G, Chaitanya Ch, Nishitha K, and Tejaswi K, IoT Based Smart Shopping Cart Integrated with Payment Gateway, in Proc. Int. Conf. Recent Advances in Electrical, Electronics, Ubiquitous Communication, and Computational Intelligence (RAEEUCCI), 2024, doi: 10.1109/RAEEUCCI61380.2024.10547971.