This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

A Survey of Challenges and Sensing Technologies in Autonomous Retail Systems

Shimmy Rukundo [email protected] 1234-5678-9012 University of MichiganAnn ArborMIUSA ,  David Wang [email protected] 1234-5678-9012 University of MichiganAnn ArborMIUSA ,  Front Wongnonthawitthaya [email protected] 734-604-1819 University of MichiganAnn ArborMIUSA ,  Youssouf Sidibé [email protected] University of MichiganAnn ArborMIUSA ,  Minsik Kim [email protected] University of MichiganAnn ArborMIUSA ,  Emily Su [email protected] University of MichiganAnn ArborMIUSA  and  Jiale Zhang University of MichiganAnn ArborMIUSA [email protected]
Abstract.

Autonomous stores leverage advanced sensing technologies to enable cashier-less shopping, real-time inventory tracking, and seamless customer interactions. However, these systems face significant challenges, including occlusion in vision-based tracking, scalability of sensor deployment, theft prevention, and real-time data processing. To address these issues, researchers have explored multi-modal sensing approaches, integrating computer vision, RFID, weight sensing, vibration-based detection, and LiDAR to enhance accuracy and efficiency. This survey provides a comprehensive review of sensing technologies used in autonomous retail environments, highlighting their strengths, limitations, and integration strategies. We categorize existing solutions across inventory tracking, environmental monitoring, people-tracking, and theft detection, discussing key challenges and emerging trends. Finally, we outline future directions for scalable, cost-efficient, and privacy-conscious autonomous store systems.

Autonomous stores, Inventory tracking, Multi-modal sensing, Computer vision, RFID Vibration, LiDAR, People-tracking, Theft detection, Sensor fusion, Real-time product detection, Environmental monitoring, AI-driven retail systems
††copyright: acmlicensed††ccs: Human-centered computing Ubiquitous and mobile computing

1. Introduction

Autonomous systems are transforming industries such as warehousing, manufacturing, and retail, enhancing efficiency and reducing operational costs. Warehouses now rely on robotic picking, autonomous guided vehicles (AGVs), and AI-driven inventory tracking to optimize logistics with minimal human intervention. Similarly, manufacturing integrates smart factories, IoT-connected machinery, and predictive maintenance to streamline production and reduce downtime. These advancements are driven by AI, real-time data processing, and sensor fusion, enabling seamless automation.

In retail, autonomous stores leverage computer vision, RFID, weight sensing, and multimodal data fusion to track inventory, monitor customer interactions, and enable cashier-less shopping. Implementations like Amazon Go use AI-powered tracking to eliminate traditional checkout, but challenges remain, including accurate inventory tracking, real-time product detection, theft prevention, and efficient environmental control.

To tackle these issues, researchers have explored various sensing and machine learning techniques. Vision-based tracking faces occlusions and product similarity issues, while weight sensors struggle with granularity for small items. Vibration-based sensing aids in shelf interaction detection and physiological monitoring, whereas RFID enhances product tracking but is costly for large-scale deployment. Additionally, people-tracking is crucial for theft detection and store optimization, with LiDAR, vibration, and RFID-based methods offering privacy-friendly alternatives to camera surveillance.

This survey examines recent advancements in inventory tracking, environmental monitoring, customer interaction detection, and theft prevention. We analyze various sensing modalities, examine their effectiveness, and discuss future directions in scalable and cost-efficient autonomous store technology.

2. Challenges

Autonomous stores aim to streamline operations by eliminating manual checkout, improving inventory tracking, and optimizing store management through sensor-driven automation. However, these systems face several technical challenges across multiple domains. Inventory monitoring remains a critical issue, as traditional vision-based tracking struggles with occlusions, similar-looking products, and dynamic customer interactions, while weight and RFID-based methods face scalability and cost constraints. Environmental conditions, such as temperature fluctuations, require precise sensor-based regulation to maintain perishable goods. Additionally, people-tracking is essential for security, theft prevention, and store layout optimization, yet existing solutions must balance accuracy, privacy concerns, and computational efficiency. Furthermore, interaction tracking—determining how customers engage with products—relies on a combination of vibration, RFID, and multimodal sensing to detect pickups, replacements, and unintended product movements. Finally, real-time operation requires fast, synchronized data processing across multiple sensing modalities, ensuring seamless inventory management, security enforcement, and customer experience in fully autonomous retail environments.

2.1. Inventory Tracking

Accurate inventory tracking is a fundamental challenge in autonomous stores, as real-time monitoring of product availability directly impacts operational efficiency, customer satisfaction, and revenue.

Vision-based systems are commonly employed for automated inventory tracking. Including in warehouses, and for stores (Kotheeranurak et al., 2023; Falcao et al., 2021; Falcão et al., 2021). These approaches face several limitations. Occlusions present a major challenge, as products may be hidden by other items, customer hands, or store layouts, making real-time recognition difficult. Additionally, similar-looking products, such as different flavors of chips or soft drinks with identical packaging, often confuse vision-based recognition models, reducing accuracy (Haritaoglu and Detection, 2001). While deep learning-based object detection (Yumang et al., 2022; Babila et al., 2022; Kanjalkar et al., 2023; Iyer, 2023; Bharadi et al., 2023; RA et al., 2024) can help mitigate these issues, these models are computationally expensive, requiring significant processing power. To improve efficiency, various studies (Kim et al., 2023b; Pestana et al., 2021; Wang and Gu, 2021; Fan et al., 2018; Kim et al., 2023a) have explored hardware acceleration techniques to speed up deep learning inference. However, scalability remains a challenge, as these models still require large, labeled datasets for training, increasing both data collection costs and deployment complexity (Ansari and Singh, 2022).

To address the shortcomings of vision-based tracking, multi-modal sensing approaches combine weight, vision, and vibration-based sensing for more accurate inventory monitoring (Ruiz et al., 2019; Falcão et al., 2020; Guinjoan et al., 2023; Lindermayr et al., 2020; Patel, 2025; Liu et al., 2024). While these multi-modal methods improve detection accuracy, they require precise synchronization and high computational power, leading to increased deployment and maintenance costs. Moreover, factors such as sensor calibration drift, noise from customer movements, and real-time processing demands pose additional challenges. Despite these difficulties, multi-modal approaches remain the most promising direction for scalable and cost-effective inventory tracking in autonomous retail environments.

2.2. Temperature Control

Maintaining optimal temperature conditions is critical for autonomous stores, especially those handling perishable goods such as fresh produce, dairy, and frozen foods. Ineffective temperature control can lead to spoilage, financial losses, and potential health hazards for consumers. Traditional refrigeration systems rely on thermostat-based regulation, but autonomous stores require more advanced, sensor-driven approaches to ensure real-time monitoring and adaptive control.

2.2.1. Sensor-Based Temperature Monitoring

Advanced temperature sensors are employed to monitor and regulate the temperature of perishable goods continuously. These sensors are integrated into an IoT-based architecture, allowing for real-time data collection, processing, and analysis to ensure optimal temperature conditions. Autonomous stores typically use wireless, low-power temperature sensors, such as MEMS-based sensors and resistance temperature detectors, to measure ambient temperature. These sensors communicate data using low-power protocols like LoRaWAN or ZigBee, enabling long-range and low-latency transmission. This allows the system to monitor temperature conditions across the entire store efficiently.

For enhanced accuracy, more specialized sensors are utilized, including thermocouples and infrared sensors. Thermocouples offer broad temperature ranges with high precision, while infrared sensors enable non-contact temperature measurements, which are particularly useful for monitoring refrigerated goods. Additionally, capacitive humidity sensors are used to measure relative humidity, a critical factor affecting the shelf life of moisture-sensitive products. Ultrasonic sensors also monitor humidity levels by detecting variations in sound wave speed caused by moisture changes. Furthermore, RFID-based monitoring systems with embedded temperature sensors are deployed to track goods throughout the supply chain, ensuring real-time logging and transmission of temperature data as items move from warehouses to store shelves.

2.2.2. Automated Temperature Regulation

Effective temperature control systems in autonomous stores combine sensor data with intelligent systems for adaptive regulation, ensuring energy efficiency while protecting perishable goods. Automated cooling systems are designed to adjust dynamically to changing environmental conditions, often using compressor-based refrigeration units with variable-speed fans and evaporators. By modulating compressor speed and fan intensity based on real-time sensor inputs, these systems maintain precise temperature control while minimizing energy consumption. The integration of machine learning models enhances the efficiency of cooling systems by predicting potential system failures and optimizing cooling strategies. These systems rely on supervised learning techniques, such as regression models or more advanced time-series forecasting models, to analyze historical temperature data and environmental conditions, adjusting operations to maintain optimal store climates and reduce energy waste.

In addition to automated cooling, AI-driven predictive maintenance plays a crucial role in maintaining the reliability of refrigeration systems. Predictive maintenance models analyze historical data and current sensor inputs to anticipate potential failures, allowing for early intervention. For instance, machine learning algorithms can forecast temperature fluctuations or system malfunctions before they occur, triggering preventive maintenance actions or adjusting refrigeration setpoints proactively. Adaptive control algorithms adjust cooling intensity based on real-time factors like customer density and external weather conditions, optimizing energy use and temperature stability. Cold chain logistics also benefit from IoT-enabled temperature tracking, with GPS systems and real-time temperature logging devices ensuring that refrigerated goods are transported under optimal conditions and deviations are promptly addressed to prevent spoilage.

Future advancements in temperature control for autonomous stores will focus on AI-driven predictive cooling, energy-efficient refrigeration, and self-healing sensor networks to improve reliability and reduce operational costs. Advanced technologies, such as solid-state cooling and quantum dot-based thermal regulation, could offer more precise and sustainable solutions. However, key challenges remain, including sensor calibration drift, network latency in real-time monitoring, and high deployment costs for advanced systems. Additionally, integrating multi-modal sensing with existing store infrastructure requires careful synchronization to prevent false alarms and inefficiencies. Addressing these challenges will be crucial for scalable, autonomous temperature management in future retail environments.

2.3. People-Tracking

Tracking people in autonomous retail environments is crucial for optimizing store layouts, enhancing loss prevention, and understanding customer behavior. Unlike traditional camera-based surveillance systems, which pose privacy risks and require high computational power, modern approaches leverage multimodal sensing to achieve accurate and privacy-preserving tracking. Methods such as LiDAR, RFID, vibration sensing, and weight-based tracking offer scalable and efficient alternatives. While vision-only approaches have been prevalent, this paper focuses on other sensing types.

Accurate people-tracking in retail stores faces several challenges:

  • •

    Occlusion & Crowd Density: Multiple customers entering together or interacting near shelves can block each other from sensor detection.

  • •

    Privacy Concerns: AI-based camera surveillance raises ethical and legal issues, leading to growing restrictions under GDPR and CCPA regulations.

  • •

    Differentiating People from Objects: Carts, strollers, or merchandise can interfere with motion-based tracking systems.

  • •

    Group Shopping Behavior: A family or group may enter together, but only one person makes a purchase, making foot traffic alone an incomplete metric for sales analysis.

To address these challenges, multimodal sensor fusion combines multiple sensing techniques to improve detection accuracy without relying solely on cameras.

LiDAR for Depth-Based Tracking

LiDAR sensors emit laser pulses to measure depth, creating a 3D map of a store’s entrance or aisles. Unlike cameras, LiDAR does not capture facial details, making it privacy-preserving while still being accurate enough to count customers.

Studies such as FAIM (Falcão et al., 2020) have explored combining weight-based tracking with vision to differentiate customer movements, reducing errors caused by occlusion. However, replacing vision-based methods with LiDAR-only solutions could offer a privacy-first alternative while still ensuring accuracy.

RFID for Shopper Identification

RFID (Radio Frequency Identification) is widely used in autonomous shopping to track product movement. When integrated with entry/exit tracking, RFID-enabled loyalty cards or wearable tags can associate individuals with purchases, solving issues related to group entry scenarios where only one person in a group buys an item.

For example, RFID exit detection systems (Bocanegra et al., 2020) scan tags at store exits, ensuring all products are accounted for before a customer leaves. However, RFID-based tracking alone cannot differentiate multiple individuals, making it less reliable for detailed people-tracking without additional sensors.

Vibration-Based Tracking

Floor vibrations can provide passive, unobtrusive tracking of people’s movement patterns (Pan et al., 2017b, 2015; Dong et al., 2022a; Mirshekari et al., 2021a, b). These works use structural vibrations from footsteps to identify customer movement, potentially allowing stores to track foot traffic patterns without the need for cameras or RFID.

Additionally, works have also explored audio-induced shelf vibrations to estimate weight changes and identify objects (Zhang et al., 2025, 2023b). This approach could be extended to detecting a person’s presence near shelves, helping track product interaction more effectively.

Sensor Fusion for Enhanced Tracking

While each sensing method has limitations, combining various sensing methods can improve and offset these issues (Pan et al., 2013; Han et al., 2018, 2017, 2018). LiDAR, RFID, and vibration sensors can improve tracking reliability. Multimodal approaches have been used in systems like VibSense (Liu et al., 2017), which leverages touch-based vibration sensing on surfaces to identify customer-product interactions.

By integrating LiDAR (depth mapping), RFID (shopper identification), and vibration sensing (movement tracking), an autonomous store can achieve:

  • •

    High-accuracy people tracking without cameras

  • •

    Group behavior differentiation

  • •

    Enhanced theft prevention with privacy compliance

2.3.1. Future Directions & Improvements

While multimodal sensing improves tracking accuracy, future research should explore:

  • •

    Machine learning models to interpret sensor fusion data more effectively.

  • •

    Energy-efficient tracking systems to reduce the computational cost of multimodal approaches.

  • •

    Combining real-time inventory tracking with foot traffic analysis for better store optimization.

3. Sensing Approaches

Autonomous stores rely on a variety of sensing technologies to track inventory, monitor customer behavior, and enhance security. Each sensing approach comes with unique advantages and challenges, making it necessary to explore multi-modal solutions that combine different sensors for improved accuracy and efficiency. This section provides an overview of key sensing modalities, including camera-based vision systems, vibration sensors, RFID tracking, LiDAR, ultrasound, and passive infrared (PIR) sensing, highlighting their roles in inventory management, theft detection, and customer interaction tracking.

3.1. Camera

Camera-based sensing is prevalent and is also a core component in autonomous settings. Camera sensors provide an advantage over existing sensors with its extensive coverage range and ability to simultaneously provide rich visual data on multiple items. Leveraging computer vision, cameras are used to perform tasks such as localization, item identification and classification, and item movement.

3.1.1. Localization and Movement

One of the main applications for cameras is precise localization in diverse environments. Studies (Dong et al., 2022b, [n. d.]; Purohit et al., 2013; Ruiz et al., 2020) have demonstrated how cameras are able to perform location tracking tailored to specific environments. Researchers are able to track the position and orientation of humans or objects using skeletal key points extracted from camera feeds. When combined with multi-view geometry (Dusmanu et al., 2020; Heng et al., 2019; Cai et al., 2019) or depth estimation (Alam et al., 2023; kim2012multifocusin; Brachmann and Rother, 2021), this visual data allows systems to calculate the relative positioning and infer movement trajectories within a known space . In autonomous stores, this system can help track customer activity and detect inventory movement. On top of solely camera-based data, current approaches explore augmenting vision-based sensors with complementary sensors to enhance temporal resolution and address privacy concerns.

3.1.2. Item Classification

Camera-based identification uses computer vision to analyze and interpret video frames. In retail environments, identification systems analyze visual components like dimensions, colors, and shapes using deep learning algorithms (Gothai et al., 2022; Franco et al., 2017). These systems can distinguish between similar-looking items with different colors schemes, providing a high accuracy rate even in crowded retail environments, but can be further refined with character recognition technology (Georgieva and Zhang, 2020; Sterkens et al., 2022; Baldominos et al., 2019; Khan et al., 2021; Fujitake, 2024). These systems pick up product labels and packaging texts from images much like the text localization techniques humans follow when given a visual data. With character recognition in place, it could provide more robust identification results that effectively spots minimal differences between physically similar retail items.

3.2. Vibration

Vibration sensors are widely used for detecting seismic events (Deng et al., 2013; Weng et al., 2011; Dean et al., 2016; Stewart et al., 2016), intrusions (Küçükbay et al., 2017; Ali et al., 2015; Hart et al., 2016; Gomery and Leach, 2000; Wang et al., 2021; Mukhopadhyay et al., 2017), and structural damage (Deraemaeker and Preumont, 2006; Panagiotopoulos et al., 2023; Yang and Han, 2002; Farrar et al., 2001). Furthermore, they have been explored for localization (Schloemann et al., 2015; Poston et al., 2015; Mirshekari et al., 2018; Poston et al., 2016; Kashimoto et al., 2016) and object tracking (Pan et al., 2016, 2017a; Drira and Smith, 2022; Poston, 2018; Pan et al., 2018; Drira et al., 2019), demonstrating their versatility across various sensing applications..

Given their versatility, vibration-based sensing has emerged as a promising approach for inventory tracking, theft detection, and even physiological monitoring in autonomous stores. Unlike vision-based or RFID-based tracking, vibration sensors can provide indirect, contactless detection of customer interactions and product movements, offering a cost-effective alternative for real-time monitoring.

3.2.1. Vibration for Inventory Monitoring

One key application of vibration sensing in autonomous stores is detecting product interactions on shelves. (Zhang et al., 2025, 2023b) leverages controlled audio-induced vibrations to monitor weight changes on shelves, and (Saidi and El Bakkali, 2023) exploits imbalanced load cells to infer product positions, allowing for real-time product detection without requiring dense load sensors. Similarly, other works (Pan et al., 2017a; Liu et al., 2017; Zhang et al., 2023d) demonstrate how surface vibrations can track taps, swipes, and even weight shifts to infer customer interactions.

3.2.2. Vibration for Theft Detection

Vibration-based sensing can enhance security in autonomous stores by detecting suspicious movements and potential shoplifting behaviors. (Pan et al., 2017b) demonstrates how ambient floor vibrations can be used to identify individual footsteps, allowing for real-time customer tracking. Similarly, (Codling et al., 2024) explores the use of indirect floor vibration sensing for heart rate and presence detection, which could help infer signs of stress or unusual behavior that may indicate theft.

Several studies have investigated the integration of vibration sensors into anti-theft systems, particularly in vehicle security (Hossain et al., 2017; Liu et al., 2013; Mamun and Ashraf, 2015; Krishnaprasad et al., 2021). Extending such techniques to high-value retail items presents a promising direction for theft prevention. However, existing systems often rely on a combination of multiple sensors, GPS trackers, and alarms, making them costly and complex for retail environments. Additionally, vehicle anti-theft systems typically monitor large, stationary objects, whereas retail products vary in size and weight, making vibration-based detection less straightforward. Furthermore, shoplifting presents a more diverse attack model, as legitimate customer interactions such as picking up, inspecting, or putting items back introduce significant ambiguity, making it difficult to distinguish between theft and normal shopping behavior. Developing a lightweight and cost-efficient vibration-based theft detection system remains an open challenge, requiring further exploration into sensor placement, signal processing, and integration with existing store security infrastructure.

3.2.3. Vibration For Product Identification

Recent work in vibration-sensing leverage sophisticated signal processing techniques and sensor fusion to enable fine-grained sensing beyond basic movement detection. These approaches (Zhang et al., 2023c, a; Chen et al., 2016; Gadre et al., 2022) leverage structural vibrations generated by products with different physical materials and characteristics. The product movements generate varying vibrational patterns, which is then used as a distinct spectral signature for classification.

3.3. RFID

RFID (radio frequency identification) is a technology that utilizes low cost wirelessly powered tags, which can be detected by a reader system. RFID tags have been incorporated into cashierless store systems (Szabó-Szentgróti et al., 2023; Al-Kassab et al., 2013; Maulana et al., 2021) for item to customer association in two main ways: exit detection based systems, where tags are scanned when a customer leaves the store, cart based systems, which rely on a RFID scanner inside a shopping cart, or tracking systems, where long range RFID scanners track the position of tags within an area.

3.3.1. Exit Based RFID Systems

Exit based RFID detection systems (Rathore and Valverde, 2011; Melià-Seguí et al., 2013) have seen varying implementations. A baseline design of an RFID checkout system (Umar et al., 2013) replaces self checkout barcode scanners with an RFID system, where users place items with attached RFID tags onto an antenna for detection. While items are identified faster than traditional barcodes, this system’s detection accuracy decreases when multiple tags are clustered together. More advanced implementations (Bocanegra et al., 2020; Busu et al., 2011) require less effort from customers to use, such as using an exit gate filled with RFID antennas that customers pass while leaving the store. Custom designed RFID readers with multiple antennas are able to scan several items at once for decreased waiting times and improved accuracy. While such approaches deliver a high level of performance, the cost of hardware is a concern for commercial implementations. (Piramuthu et al., 2014; Wen et al., 2010)

3.3.2. RFID Smart Shopping Carts

Smart shopping carts (Sawant et al., 2015; Agarwal et al., 2011; Machhirke et al., 2017) are another implementation of RFID for cashierless stores. Common smart shopping cart designs (Pradhan et al., 2017) feature a modified shopping cart with a RFID reader located at the bottom of the basket. A popular technique to ensure that only tags inside the cart are detected is through measuring the variance in RSSI. One additional benefit of measuring RSSI is that it is possible to track items the user may have interacted with outside the cart, which can provide helpful insight to store owners. Another approach (Kommavarapu et al., 2024) for item in cart detection is to scan an item’s RFID tag before it can be placed inside the cart. When an item is scanned, a door located on the cart opens, and the item can be inserted. If the user no longer wants the item, it can be scanned again to remove it from the cart’s inventory. While this approach resolves the issue of scanning multiple RFID tags in a small area, it also features lower security and greater mechanical complexity through the use of a door.

3.3.3. RFID Based Localization

RFID based position tracking and localization (Sanpechuda and Kovavisaruch, 2008; Li et al., 2019) is another idea that has been explored in cashierless store development. One aspect of localization in a cashierless store is the identification of item positions on shelves. A solution (Choi et al., 2012) is to have items with UHF RFID tags attached placed on a shelf with reference RFID tags, with a centralized RFID reader covering multiple shelves. When a new item is placed near a reference tag, an increase in interference occurs in the reference tag’s signal, which gives an estimate of the item’s position on the shelf. Item removal can be detected by comparing the tag’s RSSI to that to the typical variance in RSSI for the environment around the scanner. Room scale localization is another important component of cashierless stores. One idea for this challenge (Wang et al., 2022) is to attach RFID readers to shopping carts to track cart and customer positions. Using wall mounted boards covered in RFID tags with specific IDs, the cart based reader can detect its own position relative to the the tags. Additionally, if the tags are occluded by the shopper, the system is still able to detect the cart’s position, along with that of the shopper. Combinations of shelf and room based systems can be used to track items and shoppers across an entire store.

3.4. PIR and LiDAR

3.5. Ultrasound

Ultrasound sensing is a technique that utilizes high-frequency sound waves to detect object movements, track inventory, and enhance security in autonomous stores. Unlike vision-based systems, which can be affected by lighting conditions, or RFID-based approaches that require tag attachment, ultrasound sensing provides a contactless, infrastructure-efficient solution for monitoring product interactions and customer behavior.

Ultrasound sensors emit acoustic waves typically in the range of 20 kHz to MHz, beyond the range of human hearing. These waves propagate through the environment and reflect off objects, with the time-of-flight (ToF) and Doppler shifts of the reflected waves being analyzed to determine object distance, movement velocity, and surface characteristics. The ToF principle measures the delay between wave transmission and reception to calculate object positioning with high accuracy. Meanwhile, the Doppler effect is used to detect motion by evaluating frequency shifts in the reflected signals. More advanced implementations employ phased arrays, which manipulate wavefronts through constructive and destructive interference, enabling the formation of directional beams for improved spatial resolution. Beamforming techniques further enhance detection by filtering out unwanted reflections and focusing on specific areas of interest.

Distributed Acoustic Sensing extends ultrasound sensing principles by utilizing fiber optic cables as continuous acoustic sensors. In DAS systems, an interrogator unit transmits laser pulses through an optical fiber, and minute strain variations caused by acoustic waves induce changes in the backscattered light due to Rayleigh scattering. By analyzing these backscattered signals, DAS enables real-time detection of vibrations, movements, and disturbances along the entire length of the fiber. This technology provides a scalable and cost-effective solution for large-area monitoring without the need for discrete sensor installations. In autonomous store environments, DAS can complement ultrasound-based tracking by detecting structural vibrations associated with foot traffic, shelf interactions, and potential security breaches, further enhancing the robustness of contactless inventory and customer behavior monitoring.

4. Sensor Positioning and Distance

4.1. Optimizing Camera Placement for Comprehensive Coverage

The effectiveness of sensor technologies in autonomous retail environments is heavily influenced by their strategic positioning and distance from tracked items. According to Ruiz et al. (2019), the integration of multi-modal sensing in autonomous inventory monitoring relies critically on optimal camera placement to overcome occlusion challenges and achieve comprehensive coverage. Their research demonstrates that cameras positioned at varying heights and angles significantly improve the system’s ability to track customer-product interactions while minimizing blind spots. Similarly, Falcão et al. (2020) examined the impact of camera distance on product recognition accuracy in their FAIM framework, finding that multi-angle video collection from various distances improved the system’s ability to distinguish between similar-looking products by allowing for more precise measurement of Mahalanobis distance between product class centroids

4.2. Addressing Visual Domain Adaptation Through Sensor Positioning

Visual domain adaptation presents another critical challenge in sensor positioning for autonomous stores. AutoTag research by Ruiz et al. (2019) specifically addresses scenarios where training data is collected under different conditions than those found in actual store environments, including variations in perspective, distance, and lighting. Their findings suggest that adaptive positioning of vision sensors can mitigate these domain shift issues, with optimal results achieved when cameras can capture products from multiple perspectives that closely match those used during training. The research further indicates that combining adjustable-distance cameras with weight sensors creates a more robust system capable of adapting to varying store layouts and product arrangements. This multi-modal approach compensates for distance-related degradation in visual recognition accuracy by correlating weight measurements with visual data.

4.3. RFID Sensor Distance Calibration for Effective Product Tracking

RFID-based tracking systems also demonstrate significant sensitivity to sensor-product distances. The research by Kim et al. (2010) in their SPAMMS framework highlights how the effective range of RFID readers must be carefully calibrated based on store layout and product density. Their system dynamically adjusts tracking parameters based on distance measurements, with reader positioning optimized to maintain consistent tag detection while preventing signal interference between closely packed items. Similarly, IEEE research on RFID robots for retail item location (2025) demonstrates how mobile sensing platforms can overcome fixed-position limitations by dynamically adjusting their distance from shelved products, achieving optimal read rates at specific distance thresholds tailored to different product categories and packaging materials.

4.4. Hierarchical Sensor Deployment for Customer Interaction Monitoring

Distance considerations extend beyond product tracking to customer interaction monitoring. The AIM3S framework developed by Ruiz et al. (2019) employs strategically positioned sensors at varying distances to capture the full spectrum of customer-product interactions. Their research indicates that close-range sensors provide detailed interaction data but suffer from limited coverage, while distant sensors offer broader monitoring capabilities at the expense of granular detection. The optimal configuration involves a hierarchical deployment with distant sensors triggering more detailed analysis from closer sensors when potential interactions are detected. This tiered approach balances comprehensive coverage with detailed interaction tracking while optimizing computational resources required for real-time processing.

4.5. Adaptive Sensor Positioning for Future Autonomous Stores

Emerging research points toward adaptive sensor positioning systems that automatically reconfigure based on store conditions and monitoring requirements. Georgieva and Zhang (2020) propose optical character recognition systems for product identification that incorporate distance-adaptive focusing mechanisms, allowing for consistent performance across varying shelf depths and product placements. Their work demonstrates how autonomous calibration of sensor-to-product distances can significantly enhance recognition accuracy while reducing deployment complexity. Future autonomous store implementations will likely incorporate AI-driven optimization of sensor positioning, automatically adjusting the distance and angle of vision systems, RFID readers, and other sensors to maintain optimal coverage despite changing store layouts, product arrangements, and customer traffic patterns.

5. Conclusion

Autonomous stores rely on a combination of sensing technologies to enable cashier-less shopping, real-time inventory tracking, and enhanced security. Each sensing modality—computer vision, RFID, weight sensing, vibration-based detection, LiDAR, and ultrasound—offers unique advantages but also faces limitations in scalability, accuracy, and cost-effectiveness. To address these challenges, multi-modal sensor fusion has emerged as a promising approach, leveraging complementary sensing techniques to improve reliability and reduce deployment costs.

Despite significant advancements, open challenges remain in inventory tracking granularity, theft detection accuracy, and privacy-conscious customer monitoring. Future research should focus on energy-efficient, cost-effective, and privacy-preserving sensing systems that can seamlessly integrate into autonomous retail environments. Continued improvements in AI-driven sensor processing, real-time data fusion, and adaptive sensing techniques will be key to making autonomous stores more scalable and efficient.

References

  • (1)
  • Agarwal et al. (2011) Ankit Anil Agarwal, Saurabh Kumar Sultania, Gourav Jaiswal, and Prateek Jain. 2011. RFID based automatic shopping cart. Control Theory and Informatics 1, 1 (2011), 39–44.
  • Al-Kassab et al. (2013) Jasser Al-Kassab, Frederic Thiesse, and Thomas Buckel. 2013. RFID-enabled business process intelligence in retail stores: A case report. Journal of theoretical and applied electronic commerce research 8, 2 (2013), 112–137.
  • Alam et al. (2023) Muhammad Shamsul Alam, Farhan Bin Mohamed, Ali Selamat, and Akm Bellal Hossain. 2023. A review of recurrent neural network based camera localization for indoor environments. IEEE Access 11 (2023), 43985–44009.
  • Ali et al. (2015) T Hadj Ali, H Medjadba, LM Simohamed, and R Chemali. 2015. Intrusion detection and classification using optical fiber vibration sensor. In 2015 3rd International Conference on Control, Engineering & Information Technology (CEIT). IEEE, 1–6.
  • Ansari and Singh (2022) Mohd Aquib Ansari and Dushyant Kumar Singh. 2022. An expert video surveillance system to identify and mitigate shoplifting in megastores. Multimedia Tools and Applications 81, 16 (2022), 22497–22525.
  • Babila et al. (2022) Isaiah Francis E Babila, Shawn Anthonie E Villasor, and Jennifer C Dela Cruz. 2022. Object detection for inventory stock counting using YOLOv5. In 2022 IEEE 18th International Colloquium on Signal Processing & Applications (CSPA). IEEE, 304–309.
  • Baldominos et al. (2019) Alejandro Baldominos, Yago Saez, and Pedro Isasi. 2019. A survey of handwritten character recognition with mnist and emnist. Applied Sciences 9, 15 (2019), 3169.
  • Bharadi et al. (2023) Vinayak Bharadi, Suha Mukadam, Rahul Prasad, Kavyashree Upparakakula, and Janhavi Jaygade. 2023. Real-Time Inventory Analysis Using Jetson Nano with Object Detection and Analysis. In Deep Learning-Recent Findings and Research. IntechOpen.
  • Bocanegra et al. (2020) Carlos Bocanegra, Mohammad A Khojastepour, Mustafa Y Arslan, Eugene Chai, Sampath Rangarajan, and Kaushik R Chowdhury. 2020. RFGo: a seamless self-checkout system for apparel stores using RFID. In Proceedings of the 26th Annual International Conference on Mobile Computing and Networking. 1–14.
  • Brachmann and Rother (2021) Eric Brachmann and Carsten Rother. 2021. Visual camera re-localization from RGB and RGB-D images using DSAC. IEEE transactions on pattern analysis and machine intelligence 44, 9 (2021), 5847–5865.
  • Busu et al. (2011) MFM Busu, I Ismail, MF Saaid, and SM Norzeli. 2011. Auto-checkout system for retails using Radio Frequency Identification (RFID) technology. In 2011 IEEE Control and System Graduate Research Colloquium. IEEE, 193–196.
  • Cai et al. (2019) Ming Cai, Huangying Zhan, Chamara Saroj Weerasekera, Kejie Li, and Ian Reid. 2019. Camera relocalization by exploiting multi-view constraints for scene coordinates regression. In Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops. 0–0.
  • Chen et al. (2016) Carolyn L Chen, Jeffrey O Snyder, and Peter J Ramadge. 2016. Learning to identify container contents through tactile vibration signatures. In 2016 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR). IEEE, 43–48.
  • Choi et al. (2012) Jae Sung Choi, Hyun Lee, Daniel W. Engels, and Ramez Elmasri. 2012. Passive UHF RFID-Based Localization Using Detection of Tag Interference on Smart Shelf. IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews) 42, 2 (2012), 268–275. doi:10.1109/TSMCC.2011.2119312
  • Codling et al. (2024) Jesse R Codling, Jeffrey D Shulkin, Yen-Cheng Chang, Jiale Zhang, Hugo Latapie, Hae Young Noh, Pei Zhang, and Yiwen Dong. 2024. FloHR: Ubiquitous Heart Rate Measurement using Indirect Floor Vibration Sensing. In Proceedings of the 11th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation. 44–54.
  • Dean et al. (2016) Tim Dean, Tim Brice, Arthur Hartog, Ed Kragh, Daniele Molteni, and Kevin O’Connell. 2016. Distributed vibration sensing for seismic acquisition. The Leading Edge 35, 7 (2016), 600–604.
  • Deng et al. (2013) Tao Deng, Deyong Chen, Junbo Wang, Jian Chen, and Wentao He. 2013. A MEMS based electrochemical vibration sensor for seismic motion monitoring. Journal of Microelectromechanical Systems 23, 1 (2013), 92–99.
  • Deraemaeker and Preumont (2006) Arnaud Deraemaeker and André Preumont. 2006. Vibration based damage detection using large array sensors and spatial filters. Mechanical systems and signal processing 20, 7 (2006), 1615–1630.
  • Dong et al. (2022a) Yiwen Dong, Jonathon Fagert, Pei Zhang, and Hae Young Noh. 2022a. Stranger detection and occupant identification using structural vibrations. In European Workshop on Structural Health Monitoring. Springer, 905–914.
  • Dong et al. (2022b) Yiwen Dong, Jingxiao Liu, and Hae Young Noh. 2022b. GaitVibe+ Enhancing Structural Vibration-Based Footstep Localization Using Temporary Cameras for in-Home Gait Analysis. In Proceedings of the 20th ACM Conference on Embedded Networked Sensor Systems. 1168–1174.
  • Dong et al. ([n. d.]) Yiwen Dong, Zihao Song, Jesse Codling, Gary Rohrer, Jeremy Miles, Raj Sharma, Tami Brown-Brandl, Pei Zhang, and Hae Young Noh. [n. d.]. Robust Piglet Nursing Behavior Monitoring Through Multi-Modal Fusion of Computer Vision and Ambient Floor Vibration. Available at SSRN 4987438 ([n. d.]).
  • Drira et al. (2019) Slah Drira, Yves Reuland, Sai GS Pai, Hae Young Noh, and Ian FC Smith. 2019. Model-based occupant tracking using slab-vibration measurements. Frontiers in Built Environment 5 (2019), 63.
  • Drira and Smith (2022) Slah Drira and Ian FC Smith. 2022. A framework for occupancy detection and tracking using floor-vibration signals. Mechanical Systems and Signal Processing 168 (2022), 108472.
  • Dusmanu et al. (2020) Mihai Dusmanu, Johannes L Schönberger, and Marc Pollefeys. 2020. Multi-view optimization of local feature geometry. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part I 16. Springer, 670–686.
  • Falcão et al. (2020) João Falcão, Carlos Ruiz, Shijia Pan, Hae Young Noh, and Pei Zhang. 2020. Faim: Vision and weight sensing fusion framework for autonomous inventory monitoring in convenience stores. Frontiers in Built Environment 6 (2020), 568372.
  • Falcao et al. (2021) Joao Diogo Falcao, Prabh Simran S Baweja, Yi Wang, Akkarit Sangpetch, Hae Young Noh, Orathai Sangpetch, and Pei Zhang. 2021. Piwims: Physics informed warehouse inventory monitory via synthetic data generation. In Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers. 613–618.
  • Falcão et al. (2021) João Diogo Falcão, Carlos Ruiz, Adeola Bannis, Hae Young Noh, and Pei Zhang. 2021. Isacs: In-store autonomous checkout system for retail. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 5, 3 (2021), 1–26.
  • Fan et al. (2018) Hongxiang Fan, Shuanglong Liu, Martin Ferianc, Ho-Cheung Ng, Zhiqiang Que, Shen Liu, Xinyu Niu, and Wayne Luk. 2018. A real-time object detection accelerator with compressed SSDLite on FPGA. In 2018 International conference on field-programmable technology (FPT). IEEE, 14–21.
  • Farrar et al. (2001) Charles R Farrar, Scott W Doebling, and David A Nix. 2001. Vibration–based structural damage identification. Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences 359, 1778 (2001), 131–149.
  • Franco et al. (2017) Annalisa Franco, Davide Maltoni, and Serena Papi. 2017. Grocery product detection and recognition. Expert Systems with Applications 81 (2017), 163–176.
  • Fujitake (2024) Masato Fujitake. 2024. Dtrocr: Decoder-only transformer for optical character recognition. In Proceedings of the IEEE/CVF winter conference on applications of computer vision. 8025–8035.
  • Gadre et al. (2022) Akshay Gadre, Deepak Vasisht, Nikunj Raghuvanshi, Bodhi Priyantha, Manikanta Kotaru, Swarun Kumar, and Ranveer Chandra. 2022. MiLTOn: Sensing product integrity without opening the box using non-invasive acoustic vibrometry. In 2022 21st ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN). IEEE, 390–402.
  • Georgieva and Zhang (2020) Petia Georgieva and Pei Zhang. 2020. Optical character recognition for autonomous stores. In 2020 IEEE 10th International Conference on Intelligent Systems (IS). IEEE, 69–75.
  • Gomery and Leach (2000) Rebecca Gomery and Graham Leach. 2000. Fence vibrations [intruder detection]. IEEE Aerospace and Electronic Systems Magazine 15, 9 (2000), 3–6.
  • Gothai et al. (2022) E Gothai, Surbhi Bhatia, Aliaa Alabdali, Dilip Sharma, Bhavana Raj, and Pankaj Dadheech. 2022. Design features of grocery product recognition using deep learning. Intelligent Automation and Soft Computing 34, 2 (2022), 1231–1246.
  • Guinjoan et al. (2023) Ferran Gebellí Guinjoan, Erwin Rademakers, Abdellatif Bey Temsamani, G Radevski, T Tuytelaars, M Hutsebaut-Buysse, et al. 2023. A multi-modal ai approach for agvs: A case study on warehouse automated inventory. In Proc. of the 9th Int. Conf. on Autonomic and Autonomous Systems (ICAS), Vol. 15. 34.
  • Han et al. (2017) Jun Han, Shijia Pan, Manal Kumar Sinha, Hae Young Noh, Pei Zhang, and Patrick Tague. 2017. Sensetribute: smart home occupant identification via fusion across on-object sensing devices. In Proceedings of the 4th ACM International Conference on Systems for Energy-Efficient Built Environments. 1–10.
  • Han et al. (2018) Jun Han, Shijia Pan, Manal Kumar Sinha, Hae Young Noh, Pei Zhang, and Patrick Tague. 2018. Smart home occupant identification via sensor fusion across on-object devices. ACM Transactions on Sensor Networks (TOSN) 14, 3-4 (2018), 1–22.
  • Haritaoglu and Detection (2001) I Flickner Haritaoglu and M Detection. 2001. tracking of shopping groups in stores, Computer Vision and Pattern Recognition, 2001. CVPR 2001. In Proceedings of the 2001 IEEE Computer Society Conference on, Vol. 1.
  • Hart et al. (2016) Jan Hart, Veronika Hartova, et al. 2016. Risks of vibration detectors in intrusion and hold-up alarm systems. In 15th International scientific conference on engineering for rural development. 529.
  • Heng et al. (2019) Lionel Heng, Benjamin Choi, Zhaopeng Cui, Marcel Geppert, Sixing Hu, Benson Kuan, Peidong Liu, Rang Nguyen, Ye Chuan Yeo, Andreas Geiger, et al. 2019. Project autovision: Localization and 3d scene perception for an autonomous vehicle with a multi-camera system. In 2019 International Conference on Robotics and Automation (ICRA). IEEE, 4695–4702.
  • Hossain et al. (2017) MM Hossain, MS Islam, NF Dipu, Mohammad Tariqul Islam, Shaikh Anowarul Fattah, and Celia Shahnaz. 2017. Design of a low cost anti-theft sensor for motorcycle security device. In 2017 IEEE Region 10 Humanitarian Technology Conference (R10-HTC). IEEE, 778–783.
  • Iyer (2023) Karthik Iyer. 2023. Retail Inventory Management using Deep Learning Techniques. Ph. D. Dissertation. Dublin, National College of Ireland.
  • Kanjalkar et al. (2023) Pramod Kanjalkar, Samyak Jain, Rutuj Saraf, and Jyoti Kanjalkar. 2023. Intelligent inventory management in retail stores using four state of the art object detection models. In International Conference on Information Science and Applications. Springer, 133–148.
  • Kashimoto et al. (2016) Yukitoshi Kashimoto, Manato Fujimoto, Hirohiko Suwa, Yutaka Arakawa, and Keiichi Yasumoto. 2016. Floor vibration type estimation with piezo sensor toward indoor positioning system. In 2016 International Conference on Indoor Positioning and Indoor Navigation (IPIN). IEEE, 1–6.
  • Khan et al. (2021) Abdullah Ayub Khan, Asif Ali Laghari, and Shafique Ahmed Awan. 2021. Machine learning in computer vision: A review. EAI Endorsed Transactions on Scalable Information Systems 8, 32 (2021).
  • Kim et al. (2023a) Kyungho Kim, Sung-Joon Jang, Jonghee Park, Eunchong Lee, and Sang-Seol Lee. 2023a. Lightweight and energy-efficient deep learning accelerator for real-time object detection on edge devices. Sensors 23, 3 (2023), 1185.
  • Kim et al. (2023b) Minsik Kim, Kyoungseok Oh, Youngmock Cho, Hojin Seo, Xuan Truong Nguyen, and Hyuk-Jae Lee. 2023b. A low-latency FPGA accelerator for YOLOv3-tiny with flexible layerwise mapping and dataflow. IEEE Transactions on Circuits and Systems I: Regular Papers 71, 3 (2023), 1158–1171.
  • Kommavarapu et al. (2024) Lohith Kommavarapu, Thanmayee Mangu, Suraj Preetham Macherla, and Ball. Mukund. Mani Tripathi. 2024. Smart Trolley for Quick Shopping. In 2024 IEEE International Conference on Computing, Power and Communication Technologies (IC2PCT), Vol. 5. 141–144. doi:10.1109/IC2PCT60090.2024.10486478
  • Kotheeranurak et al. (2023) Dechabhol Kotheeranurak, Orathai Sangpetch, Hae Young Noh, and Pei Zhang. 2023. Long-Term Inventory Flow Management System Tracking Using Doorway Monitoring. In Proceedings of the 22nd International Conference on Information Processing in Sensor Networks. 324–325.
  • Krishnaprasad et al. (2021) C Krishnaprasad, CR Albin Joseph, IS Sarath, and O Rahul Manohar. 2021. A novel low-cost theft detection system for two wheelers with minimum carbon foot print. In 2021 2nd International Conference for Emerging Technology (INCET). IEEE, 1–5.
  • Küçükbay et al. (2017) Selver Ezgi Küçükbay, Mustafa Sert, and Adnan Yazici. 2017. Use of acoustic and vibration sensor data to detect objects in surveillance wireless sensor networks. In 2017 21st International Conference on Control Systems and Computer Science (CSCS). IEEE, 207–212.
  • Li et al. (2019) Chenyang Li, Lingfei Mo, and Dongkai Zhang. 2019. Review on UHF RFID localization methods. IEEE Journal of Radio Frequency Identification 3, 4 (2019), 205–215.
  • Lindermayr et al. (2020) Jochen Lindermayr, Cagatay Odabasi, Timur Wohlleber, and Birgit Graf. 2020. Multi-modal Visual Withdrawal Detection for Inventory Management on a Robotic Care Cart. In ISR 2020; 52th International Symposium on Robotics. VDE, 1–6.
  • Liu et al. (2017) Jian Liu, Yingying Chen, Marco Gruteser, and Yan Wang. 2017. Vibsense: Sensing touches on ubiquitous surfaces through vibration. In 2017 14th Annual IEEE International Conference on Sensing, Communication, and Networking (SECON). IEEE, 1–9.
  • Liu et al. (2024) Jingya Liu, Issac Huang, Aishwarya Anand, Po-Hao Chang, and Yufang Huang. 2024. Digital Twin in Retail: An AI-Driven Multi-Modal Approach for Real-Time Product Recognition and 3D Store Reconstruction. In 2024 IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW). IEEE, 368–373.
  • Liu et al. (2013) Zhigang Liu, Anqi Zhang, and Shaojun Li. 2013. Vehicle anti-theft tracking system based on Internet of things. In Proceedings of 2013 IEEE International Conference on Vehicular Electronics and Safety. IEEE, 48–52.
  • Machhirke et al. (2017) Komal Machhirke, Priyanka Goche, Rupali Rathod, Rinku Petkar, and Manohar Golait. 2017. A new technology of smart shopping cart using RFID and ZigBee. International Journal on Recent and Innovation Trends in Computing and Communication 5, 2 (2017), 256–259.
  • Mamun and Ashraf (2015) Kabir A Mamun and Zahir Ashraf. 2015. Anti-theft vehicle security system with preventive action. In 2015 2nd Asia-Pacific World Congress on Computer Science and Engineering (APWC on CSE). IEEE, 1–6.
  • Maulana et al. (2021) Fachrurrozi Maulana, Rizky Prawira Putra, Novita Hanafiah, et al. 2021. Self-checkout system using RFID (Radio Frequency Identification) technology: A survey. In 2021 1st International Conference on Computer Science and Artificial Intelligence (ICCSAI), Vol. 1. IEEE, 273–277.
  • Melià-Seguí et al. (2013) Joan Melià-Seguí, Rafael Pous, Anna Carreras, Marc Morenza-Cinos, Raúl Parada, Zeinab Liaghat, and Ramir De Porrata-Doria. 2013. Enhancing the shopping experience through RFID in an actual retail store. In Proceedings of the 2013 ACM conference on Pervasive and ubiquitous computing adjunct publication. 1029–1036.
  • Mirshekari et al. (2021a) Mostafa Mirshekari, Jonathon Fagert, Shijia Pan, Pei Zhang, and Hae Young Noh. 2021a. Obstruction-invariant occupant localization using footstep-induced structural vibrations. Mechanical Systems and Signal Processing 153 (2021), 107499.
  • Mirshekari et al. (2021b) Mostafa Mirshekari, Jonathon Fagert, Shijia Pan, Pei Zhang, and Hae Young Noh. 2021b. Occupant Localization in Obstructive Indoor Environments Using Footstep-Induced Floor Vibrations. In Dynamics of Civil Structures, Volume 2: Proceedings of the 38th IMAC, A Conference and Exposition on Structural Dynamics 2020. Springer, 121–124.
  • Mirshekari et al. (2018) Mostafa Mirshekari, Shijia Pan, Jonathon Fagert, Eve M Schooler, Pei Zhang, and Hae Young Noh. 2018. Occupant localization using footstep-induced structural vibration. Mechanical Systems and Signal Processing 112 (2018), 77–97.
  • Mukhopadhyay et al. (2017) Bodhibrata Mukhopadhyay, Sahil Anchal, and Subrat Kar. 2017. Detection of an intruder and prediction of his state of motion by using seismic sensor. IEEE Sensors Journal 18, 2 (2017), 703–712.
  • Pan et al. (2013) Shijia Pan, An Chen, and Pei Zhang. 2013. Securitas: user identification through rgb-nir camera pair on mobile devices. In Proceedings of the Third ACM workshop on Security and privacy in smartphones & mobile devices. 99–104.
  • Pan et al. (2016) Shijia Pan, Kent Lyons, Mostafa Mirshekari, Hae Young Noh, and Pei Zhang. 2016. Multiple pedestrian tracking through ambient structural vibration sensing. In Proceedings of the 14th ACM Conference on Embedded Network Sensor Systems CD-ROM. 366–367.
  • Pan et al. (2018) Shijia Pan, Mostafa Mirshekari, Jonathon Fagert, Ceferino Gabriel Ramirez, Albert Jin Chung, Chih Chi Hu, John Paul Shen, Pei Zhang, and Hae Young Noh. 2018. Characterizing human activity induced impulse and slip-pulse excitations through structural vibration. Journal of Sound and Vibration 414 (2018), 61–80.
  • Pan et al. (2017a) Shijia Pan, Ceferino Gabriel Ramirez, Mostafa Mirshekari, Jonathon Fagert, Albert Jin Chung, Chih Chi Hu, John Paul Shen, Hae Young Noh, and Pei Zhang. 2017a. Surfacevibe: vibration-based tap & swipe tracking on ubiquitous surfaces. In Proceedings of the 16th ACM/IEEE International Conference on Information Processing in Sensor Networks. 197–208.
  • Pan et al. (2015) Shijia Pan, Ningning Wang, Yuqiu Qian, Irem Velibeyoglu, Hae Young Noh, and Pei Zhang. 2015. Indoor person identification through footstep induced structural vibration. In Proceedings of the 16th International Workshop on Mobile Computing Systems and Applications. 81–86.
  • Pan et al. (2017b) Shijia Pan, Tong Yu, Mostafa Mirshekari, Jonathon Fagert, Amelie Bonde, Ole J Mengshoel, Hae Young Noh, and Pei Zhang. 2017b. Footprintid: Indoor pedestrian identification through ambient structural vibration sensing. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 1, 3 (2017), 1–31.
  • Panagiotopoulos et al. (2023) Andeas Panagiotopoulos, Tcherniak Dmitri, and Fassois D Spilios. 2023. Damage detection on the blade of an operating wind turbine via a single vibration sensor and statistical time series methods: Exploring the performance limits of robust methods. Structural Health Monitoring 22, 1 (2023), 433–448.
  • Patel (2025) S Patel. 2025. Multi-Modal product recognition in retail environments: Enhancing accuracy through integrated vision and OCR approaches. World Journal of Advanced Research and Reviews 25, 1 (2025), 1837–1844.
  • Pestana et al. (2021) Daniel Pestana, Pedro R Miranda, João D Lopes, Rui P Duarte, Mário P Véstias, Horácio C Neto, and José T De Sousa. 2021. A full featured configurable accelerator for object detection with YOLO. IEEE Access 9 (2021), 75864–75877.
  • Piramuthu et al. (2014) Selwyn Piramuthu, Sina Wochner, and Martin Grunow. 2014. Should retail stores also RFID-tag ‘cheap’items? European Journal of Operational Research 233, 1 (2014), 281–291.
  • Poston (2018) Jeffrey D Poston. 2018. Toward tracking multiple building occupants by footstep vibrations. In 2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP). IEEE, 86–90.
  • Poston et al. (2016) Jeffrey D Poston, R Michael Buehrer, Americo G Woolard, and Pablo A Tarazaga. 2016. Indoor positioning from vibration localization in smart buildings. In 2016 IEEE/ION Position, Location and Navigation Symposium (PLANS). IEEE, 366–372.
  • Poston et al. (2015) Jeffrey D Poston, Javier Schloemann, R Michael Buehrer, VVN Sriram Malladi, Americo G Woolard, and Pablo A Tarazaga. 2015. Towards indoor localization of pedestrians via smart building vibration sensing. In 2015 international conference on localization and GNSS (ICL-GNSS). IEEE, 1–6.
  • Pradhan et al. (2017) Swadhin Pradhan, Eugene Chai, Karthik Sundaresan, Sampath Rangarajan, and Lili Qiu. 2017. Konark: A RFID based system for enhancing in-store shopping experience. In Proceedings of the 4th International on Workshop on Physical Analytics. 19–24.
  • Purohit et al. (2013) Aveek Purohit, Zheng Sun, Shijia Pan, and Pei Zhang. 2013. SugarTrail: Indoor navigation in retail environments without surveys and maps. In 2013 IEEE International Conference on Sensing, Communications and Networking (SECON). IEEE, 300–308.
  • RA et al. (2024) Vijeya Sarathe K RA, G Jaspher W Kathrine, and S Salaja. 2024. Dynamic Inventory Object Detection Using YOLOv8 Deep Learning Model. In 2024 International Conference on Advances in Modern Age Technologies for Health and Engineering Science (AMATHE). IEEE, 1–6.
  • Rathore and Valverde (2011) Athul Rathore and Raul Valverde. 2011. An RFID based E-commerce solution for the implementation of secure unattended stores. Journal of Emerging Trends in Computing and Information Sciences 2, 8 (2011), 376–389.
  • Ruiz et al. (2019) Carlos Ruiz, Joao Falcao, Shijia Pan, Hae Young Noh, and Pei Zhang. 2019. Aim3s: Autonomous inventory monitoring through multi-modal sensing for cashier-less convenience stores. In Proceedings of the 6th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation. 135–144.
  • Ruiz et al. (2020) Carlos Ruiz, Shijia Pan, Adeola Bannis, Ming-Po Chang, Hae Young Noh, and Pei Zhang. 2020. IDIoT: Towards ubiquitous identification of IoT devices through visual and inertial orientation matching during human activity. In 2020 IEEE/ACM Fifth International Conference on Internet-of-Things Design and Implementation (IoTDI). IEEE, 40–52.
  • Saidi and El Bakkali (2023) Tarik Saidi and Abdelmajid El Bakkali. 2023. Smart position detection of shelf-load based on imbalanced load cell sensors. In 2023 IEEE International Conference on Advances in Data-Driven Analytics And Intelligent Systems (ADACIS). IEEE, 1–5.
  • Sanpechuda and Kovavisaruch (2008) Taweesak Sanpechuda and La-or Kovavisaruch. 2008. A review of RFID localization: Applications and techniques. In 2008 5th international conference on electrical engineering/electronics, computer, telecommunications and information technology, Vol. 2. IEEE, 769–772.
  • Sawant et al. (2015) Ms Rupali Sawant, Kripa Krishnan, Shweta Bhokre, and Priyanka Bhosale. 2015. The RFID based smart shopping cart. International Journal of Engineering Research and General Science 3, 2 (2015), 275–280.
  • Schloemann et al. (2015) Javier Schloemann, VVN Sriram Malladi, Americo G Woolard, Joseph M Hamilton, R Michael Buehrer, and Pablo A Tarazaga. 2015. Vibration event localization in an instrumented building. In Experimental Techniques, Rotating Machinery, and Acoustics, Volume 8: Proceedings of the 33rd IMAC, A Conference and Exposition on Structural Dynamics, 2015. Springer, 265–271.
  • Sterkens et al. (2022) Wouter Sterkens, Dillam Jossue Diaz-Romero, Benya Xiao, Bowen Wu, Toon Goedemé, Wim Dewulf, and Jef Peeters. 2022. Computer vision and optical character recognition for the classification of batteries from WEEE. Procedia CIRP 105 (2022), 110–115.
  • Stewart et al. (2016) Robert Stewart, Li Chang, Srikanth Sudarshan, Aaron Becker, and Li Huang. 2016. An unmanned aerial vehicle with vibration sensing ability (seismic drone). In SEG International Exposition and Annual Meeting. SEG, SEG–2016.
  • Szabó-Szentgróti et al. (2023) Eszter Szabó-Szentgróti, Szabolcs Rámháp, and Petra Kinga Kézai. 2023. Systematic review of cashierless stores (just walk out stores) revolutionizing the retail. Management & Marketing 18, s1 (2023), 427–448.
  • Umar et al. (2013) Mohammad Sarosh Umar, Juned Ahmad Ansari, and Mohammad Qasim Rafiq. 2013. Automated Retail Store Based on RFID. In 2013 International Conference on Communication Systems and Network Technologies. 17–21. doi:10.1109/CSNT.2013.12
  • Wang and Gu (2021) Jin Wang and Shenshen Gu. 2021. Fpga implementation of object detection accelerator based on vitis-ai. In 2021 11th International Conference on Information Science and Technology (ICIST). IEEE, 571–577.
  • Wang et al. (2021) Jie Wang, Yang Qiu, and Sheng Li. 2021. Intruder detection trial in subway tunnel based on distributed vibration response. In IOP Conference Series: Earth and Environmental Science, Vol. 638. IOP Publishing, 012080.
  • Wang et al. (2022) Pei Wang, Bin Guo, Zhu Wang, and Zhiwen Yu. 2022. ShopSense:Customer Localization in Multi-Person Scenario With Passive RFID Tags. IEEE Transactions on Mobile Computing 21, 5 (2022), 1812–1828. doi:10.1109/TMC.2020.3029833
  • Wen et al. (2010) Ta-Chun Wen, Yung-Chia Chang, and Kuei-Hu Chang. 2010. Cost-benefit analysis of RFID application in apparel retailing for SME: A case from Taiwan. Transportation Journal 49, 3 (2010), 57–66.
  • Weng et al. (2011) Yinyan Weng, Xueguang Qiao, Tuan Guo, Manli Hu, Zhongyao Feng, Ruohui Wang, and Jing Zhang. 2011. A robust and compact fiber Bragg grating vibration sensor for seismic measurement. IEEE sensors journal 12, 4 (2011), 800–804.
  • Yang and Han (2002) YC Yang and KS Han. 2002. Damage monitoring and impact detectionusing optical fiber vibration sensors. Smart materials and structures 11, 3 (2002), 337.
  • Yumang et al. (2022) Analyn N Yumang, John Anthony S Katipunan, and Abigail L Galido. 2022. Recognition and Counting of Materials Used in Fast Food Restaurants Based on YOLO Object Detection for Inventory System. In 2022 IEEE 14th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM). IEEE, 1–6.
  • Zhang et al. (2023b) Jiale Zhang, Shweta Pati, Jesse Codling, Adeola Bannis, Carlos Ruiz, Hae Young Noh, and Pei Zhang. 2023b. Vibration-Based Object Classification with Structural Response of Ambient Music. In Proceedings of the 22nd International Conference on Information Processing in Sensor Networks. 314–315.
  • Zhang et al. (2025) Jiale Zhang, Yuyan Wu, Jesse R Codling, Yen Cheng Chang, Julia Gersey, Pei Zhang, Hae Young Noh, and Yiwen Dong. 2025. WeVibe: Weight Change Estimation Through Audio-Induced Shelf Vibrations In Autonomous Stores. arXiv preprint arXiv:2502.12093 (2025).
  • Zhang et al. (2023a) Yue Zhang, Shiwei Fang, Carlos Ruiz, Zhizhang Hu, Shubham Rohal, and Shijia Pan. 2023a. Augmenting Vibration-Based Customer-Product Interaction Recognition with Sparse Load Sensing. In Proceedings of Cyber-Physical Systems and Internet of Things Week 2023. 266–271.
  • Zhang et al. (2023c) Yue Zhang, Carlos Ruiz, Shubham Rohal, and Shijia Pan. 2023c. CPA: Cyber-Physical Augmentation for Vibration Sensing in Autonomous Retails. In Proceedings of the 24th International Workshop on Mobile Computing Systems and Applications. 8–14.
  • Zhang et al. (2023d) Yue Zhang, Carlos Ruiz, Shubham Rohal, and Shijia Pan. 2023d. Single-Point Vibration Sensing for Product Pickup/Put-Down Detection in Autonomous Retails. In Proceedings of the 24th International Workshop on Mobile Computing Systems and Applications. 136–136.