Proximity Video Interaction Detection for Non-Discretized Items
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensors in materials handling facilities are unable to detect interactions with non-discretized items, such as liquids or solids sold in varying quantities and container sizes, due to their inability to determine measurable quantities associated with these interactions.
Innovation Solution
The system captures digital imagery from within the materials handling facility and defines product spaces associated with dispensing units. By tracking customer locations and determining if they breach or are within a predetermined range of these product spaces, the system confirms interactions through video feeds evaluated by human associates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sensors (imaging devices, load sensors, RFID sensors) are used to detect interactions with discretized items, then detection accuracy for discrete items is improved, but the system becomes unable to detect interactions with non-discretized items
Solution Approach 1:
The patent introduces an intermediary computational model (physics-based simulation model) that translates sensor observations into meaningful interactions. This intermediary layer processes raw sensor data through simulated physics engines to infer interactions with non-discretized items, enabling the system to handle both discrete and continuous items uniformly without requiring different sensor types
2Loss of information
If sensors are configured to detect discretized items using mass changes, RFID signals, or image recognition, then item identification is improved, but the system cannot determine measurable quantities for non-discretized items
Solution Approach 1:
The patent changes the parameter representation from discrete item attributes (RFID tags, binary presence/absence) to continuous physical parameters (mass, volume, density). By modeling items as physical substances with measurable parameters and using physics-based simulations, the system can determine quantities of non-discretized items by tracking changes in these continuous parameters rather than relying on discrete identification signals
3Loss of information
If imaging devices are used to capture digital imagery for interaction detection, then visual identification of items is improved, but the system becomes ill-suited for calculating mass, volume, or other measurable quantities
Solution Approach 1:
The patent makes the imaging device universal by integrating it with a physics-based computational model that enables multiple functions: visual identification, mass calculation, volume determination, and interaction detection. The same image data is processed through physics simulations to extract various physical quantities, allowing a single sensor type to perform both qualitative and quantitative measurements
Data Source
AI summary
Product spaces for containers of non-discretized items are defined by sets of points in three-dimensional space. When a customer breaches or contacts a product space established for a container of the non-discretized items, video feeds representing the product space are displayed to an associate in a web page or another user interface. By interacting with the user interface, the associate may view the video feeds and make a determination as to whether the customer received any of the non-discretized items from the container, and to confirm an identity of the customer accordingly.