Robotic Item Identification Using Multimodal Sensor Fusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robotic item manipulation systems rely solely on barcodes for identification, which can lead to errors due to obscured or damaged codes, and are inefficient in handling items with similar features or unstructured arrangements.

Innovation Solution

A multimodal identification system using multiple sensors to detect characteristics such as visual appearance, weight, and dimension, allowing for identification even when codes cannot be scanned, and validating or supplementing barcode data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If barcode scanning is used for item identification, then the identification process is simple and fast, but the system reliability deteriorates when codes are obscured or damaged

Engineering Contradiction:
Improveidentification speedVSAvoididentification reliability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent combines multiple identification methods (barcode scanning, visual appearance recognition, weight measurement, dimension detection) into a unified identification system. The sensor array captures multiple characteristics simultaneously, and the system merges these data sources to identify items, ensuring reliable identification even when barcode scanning fails due to obscured or damaged codes.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces an intermediary identification layer that uses visual appearance, weight, and dimension characteristics as intermediate data sources when primary barcode identification fails. This intermediary system acts as a backup and validation mechanism, transferring identification capability from the failed barcode method to alternative characteristic-based methods.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If multiple sensor modalities are used for item identification, then the identification reliability improves, but the device complexity increases

Engineering Contradiction:
Improveidentification reliabilityVSAvoidsensor system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent implements a multi-functional sensor array where sensors serve multiple purposes: capturing visual appearance for identification, detecting weight for verification, and measuring dimensions for validation. This universal sensor system handles multiple identification tasks simultaneously, reducing the need for separate specialized devices for each sensing modality.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent segments the identification process into distinct characteristic detection modules (visual appearance detection, weight detection, dimension detection) that operate independently but contribute to the overall identification. This modular segmentation allows each sensor type to be optimized for its specific function while maintaining system manageability and reducing overall complexity.

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If barcode-only identification is used, then the system is easy to operate, but the productivity decreases when items have similar features or unstructured arrangements

Engineering Contradiction:
Improvesystem operation simplicityVSAvoiditem handling productivity
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent performs preliminary detection of multiple item characteristics (visual appearance, weight, dimension) before the identification decision is made. By capturing all relevant data in advance through the sensor array, the system prepares comprehensive identification information that enables rapid and accurate item differentiation, even for items with similar features or in unstructured arrangements, thereby improving productivity without complicating operation.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS11970343B1Multimodal identification system and method for robotic item manipulators
Publication Date: 2024.04.30 AMAZON TECH INC
  • US11970343B1 patent drawing
  • US11970343B1 patent drawing
  • US11970343B1 patent drawing

AI summary

An apparatus includes a memory and a hardware processor. The hardware processor determines one or more of a weight of an item, a packaging type of the item, a packaging material of the item, a barcode of the item, a rigidity of the item, or a physical response of the item to being lifted and determines a visual appearance of the item and a shape or size of the item. The hardware processor also compares, using a machine learning model, the determined characteristics of the item to a manifest for the container. The manifest identifies a plurality of items in the container. The hardware processor determines, using the machine learning model, an identity of the item based on comparing the determined characteristics of the item to the manifest.