A Survey of Challenges and Sensing Technologies in Autonomous Retail Systems
Shimmy Rukundo, David Wang, Front Wongnonthawitthaya, Youssouf Sidibé, Minsik Kim, Emily Su, Jiale Zhang
TL;DR
This survey addresses the challenges of cashier-less autonomous retail by examining a range of sensing modalities—vision, RFID, weight, vibration, LiDAR, PIR, and ultrasound—and their roles in inventory tracking, environmental monitoring, people-tracking, and theft prevention. It emphasizes multi-modal sensor fusion as a path to improved accuracy, privacy preservation, and scalability, detailing strengths, limitations, and integration strategies across inventory and customer-interaction contexts. The paper identifies critical issues such as occlusion, perishable goods temperature control, real-time data processing, and privacy concerns, and proposes directions toward energy-efficient, cost-effective, and privacy-conscious sensing systems with AI-driven sensor positioning and adaptive configurations. Overall, the work provides a structured synthesis of current technologies and outlines practical avenues for building robust, scalable autonomous retail environments.
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.