AGV Inventory Storage Using Visual Dissimilarity Models

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing inventory management systems face inefficiencies in recognizing and picking items from containers, as they often require significant effort and resources, and lead to unbalanced weight distribution and wasted storage space due to separate compartments for each item type.

Innovation Solution

A method using a trained dissimilarity model to determine which item types can be stored together in a container based on their appearance features, allowing for efficient item recognition and improved container stability by eliminating the need for multiple compartments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If different types of items are stored in separate containers or partitions, then item recognition efficiency is improved, but storage space utilization deteriorates

Engineering Contradiction:

Improveitem recognition efficiencyVSAvoidstorage space utilization

Core Design Contradiction:

ProductivityVSVolume of stationary object

Solution Approach 1:

The system segments items based on their visual features and stores them in the same container when they have sufficient visual dissimilarity. The dissimilarity model segments the feature space rather than physically separating items, allowing multiple item types to coexist in one container while maintaining recognition efficiency.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent merges multiple item types into the same container by evaluating their visual dissimilarity scores. Items with high dissimilarity scores are combined in the same container, maximizing storage space utilization while ensuring they remain distinguishable through their visual characteristics.

Inventive Principle:

Principle #5Merging (Combining)

2Ease of operation

If items of one type are stored in the same compartment, then item recognition is simplified, but weight distribution balance deteriorates

Engineering Contradiction:

Improveitem recognition simplicityVSAvoidcontainer weight distribution

Core Design Contradiction:

Ease of operationVSStability of the object's composition

Solution Approach 1:

The system applies local quality by assigning different item types to different spatial locations within the container based on their visual features and weight characteristics. This ensures both recognition simplicity and weight balance by strategically positioning items rather than uniformly distributing them.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent employs dynamic placement strategies that adapt to the specific combination of items in each container. The system dynamically adjusts item positions and container assignments based on real-time calculations of weight distribution and visual dissimilarity, rather than following fixed static rules.

Inventive Principle:

Principle #15Dynamics

3Volume of stationary object

If multiple item types are stored together in a container, then storage space utilization is improved, but item recognition complexity increases

Engineering Contradiction:

Improvestorage space utilizationVSAvoiditem recognition complexity

Core Design Contradiction:

Volume of stationary objectVSDevice complexity

Solution Approach 1:

The dissimilarity model acts as an intermediary that pre-evaluates and scores the visual differences between item types before they are stored together. This intermediary assessment allows the system to confidently store multiple item types in one container without increasing recognition complexity, as the model has already verified their distinguishability.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary action by calculating dissimilarity scores and evaluating item compatibility before actual storage. This advance assessment ensures that only item combinations with sufficient visual distinction are placed together, preventing recognition complexity issues from arising during the picking process.

Inventive Principle:

Principle #10Preliminary action

4Productivity

If a dissimilarity model is used to determine item placement, then storage optimization is improved, but computational resources increase

Engineering Contradiction:

Improvestorage optimizationVSAvoidcomputational resource consumption

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The system applies partial action by using the dissimilarity model selectively for item pairs that benefit most from co-storage. Rather than exhaustively analyzing all possible item combinations, the model focuses computational resources on evaluating promising candidates, achieving storage optimization with reduced computational overhead.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11498776B1Automated guided vehicle control and organizing inventory items using dissimilarity models

Publication Date: 2022.11.15 STAPLES INC

AI summary

In an embodiment, a method may determine a first candidate item type and a second candidate item type for a container. The method may determine appearance feature(s) of the first candidate item type and appearance feature(s) of the second candidate item type. The method may determine, using a trained dissimilarity model, a distinguishability score between the first candidate item type and the second candidate item type based on the appearance feature(s) of the first candidate item type and the appearance feature(s) of the second candidate item type. The method may determine to store the first candidate item type and the second candidate item type together in the container based on the distinguishability score between the first candidate item type and the second candidate item type. In some instances, the method may place item(s) of the first candidate item type and item(s) of the second candidate item type in the container.