Retail Checkout Sensing With Confidence-Based Virtual Carts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional physical stores require manual checkout processes for customers to purchase items, which can be time-consuming and prone to errors or fraud, and existing automated checkout systems lack real-time confidence in item selection and tracking.
Innovation Solution
A system utilizing facility sensors and mobile apparatuses to track user and item locations, enabling automated checkout by associating item identifiers with user accounts and processing transactions without manual intervention, while providing real-time cart content tracking and confidence scoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual checkout processes are used in traditional physical stores, then customers can purchase items with human assistance and verification, but the process becomes time-consuming and prone to errors or fraud
Solution Approach 1:
The system enables automated checkout where customers serve themselves by picking items and exiting without manual cashier intervention. Sensors automatically detect items taken, track them in a virtual cart, and process payments without human assistance, eliminating checkout wait times while maintaining reliability through multiple verification mechanisms.
Solution Approach 2:
The patent replaces manual mechanical checkout processes with an automated sensor-based system. Cameras, weight sensors, and RFID readers detect and track items, while computer vision algorithms and confidence scoring systems process transactions electronically, substituting human cashier operations with automated technological systems.
2Productivity
If automated checkout systems are implemented to reduce manual intervention, then checkout speed increases, but real-time confidence in item selection and tracking becomes insufficient
Solution Approach 1:
The system continuously monitors item selection events and updates confidence scores in real-time based on sensor data from multiple sources. This feedback mechanism allows the system to maintain high confidence in automated detection by constantly verifying item locations, cart contents, and customer actions across different sensor inputs.
Solution Approach 2:
The patent employs multiple sensor types (cameras, weight sensors, RFID readers) that serve multiple functions: detecting item pickup, tracking item location, verifying cart contents, and preventing fraud. This multi-functional approach enhances measurement precision by cross-validating data from different sensor sources simultaneously.
3Reliability
If sensors and mobile apparatuses are deployed throughout the facility to track users and items, then automated checkout accuracy improves, but system complexity and cost increase
Solution Approach 1:
Each sensor and mobile apparatus is designed to perform multiple functions simultaneously. Cameras detect item pickup, track customer movement, and verify checkout completion. Weight sensors monitor cart weight changes to detect item addition/removal. This multi-functionality reduces the total number of components needed while maintaining high tracking accuracy.
Solution Approach 2:
The patent combines multiple sensing capabilities into integrated mobile apparatuses and fixed sensor systems. Rather than deploying separate devices for each function, the system merges camera, RFID, and weight sensing capabilities into unified tracking infrastructure that serves multiple purposes across the facility.
4Object-generated harmful factors
If real-time cart content tracking is implemented with confidence scoring, then fraud and shrink are reduced, but data processing requirements and computational load increase
Solution Approach 1:
The system processes data at different levels of detail based on confidence thresholds. For high-confidence events, minimal processing is required. For lower-confidence situations, the system applies more intensive verification only when necessary, rather than uniformly processing all data at maximum detail, thus reducing overall computational energy consumption.
Solution Approach 2:
The system pre-processes and validates sensor data as it is collected, performing initial confidence assessments and filtering before full transaction processing. This preliminary action reduces the computational load on downstream systems by eliminating obviously valid or invalid events before they require intensive processing.
Data Source
AI summary
This disclosure describes, in part, systems for enabling facilities to implement techniques to determine when users are in possession of items when located within and/or exiting the facilities. For instance, a system may use one or more sensors of a facility and of a mobile apparatus to determine interactions with items in a facility. The system may determine confidence scores for the interactions and may direct the user to varying checkout experiences based on the confidence scores associated with virtual carts of their items and may enable additional user experiences and interactions with the facility and items of the facility to improve customer experiences.