Image Sensor Object Tracking Reducing Radar Weight

Solution Overview

Problem

Existing image sensor systems for object tracking and classification are often bulky and heavy due to the need for radar systems, and they require significant computational resources and data storage, which increases size, weight, and power (SWaP) requirements, making them unsuitable for smaller vehicles.

Innovation Solution

The use of an image sensor system with interface circuitry, machine-readable instructions, and programmable circuitry to preprocess images, identify potential targets, filter images to determine persistent objects, characterize looming characteristics, and classify objects based on these characteristics, thereby reducing the need for bulky radar systems and minimizing SWaP requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If radar systems are used for object tracking and classification, then detection reliability is improved, but device weight and size increase

Engineering Contradiction:

Improvedetection reliabilityVSAvoidsystem weight

Core Design Contradiction:

ReliabilityVSWeight of moving object

Solution Approach 1:

The patent replaces the mechanical/radar-based detection system with an optical image sensor system combined with machine learning algorithms. The image sensor captures visual data that is processed through trained neural networks to identify and classify objects, substituting the need for heavy radar hardware while maintaining detection capabilities through intelligent image analysis

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system creates a digital copy of the detection function by training machine learning models on image data. Instead of using physical radar sensors to detect objects, the system captures images and creates computational representations (embeddings) of objects through trained neural networks, allowing object detection without the physical radar hardware

Inventive Principle:

Principle #26Copying

2Reliability

If radar systems are used for object tracking and classification, then detection reliability is improved, but device volume increases

Engineering Contradiction:

Improvedetection reliabilityVSAvoidsystem volume

Core Design Contradiction:

ReliabilityVSVolume of moving object

Solution Approach 1:

The patent replaces the mechanical/radar-based detection system with an optical image sensor system combined with machine learning algorithms. The image sensor captures visual data that is processed through trained neural networks to identify and classify objects, substituting the need for heavy radar hardware while maintaining detection capabilities through intelligent image analysis

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system creates a digital copy of the detection function by training machine learning models on image data. Instead of using physical radar sensors to detect objects, the system captures images and creates computational representations (embeddings) of objects through trained neural networks, allowing object detection without the physical radar hardware

Inventive Principle:

Principle #26Copying

3Measurement precision

If image sensor systems process more images for better object identification, then measurement precision is improved, but computational resources and data storage requirements increase

Engineering Contradiction:

Improveobject identification precisionVSAvoidcomputational resources

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary action by pre-training machine learning models on large datasets of images before actual operation. The neural networks are trained in advance to recognize object patterns, features, and characteristics, so that during real-time operation the system only needs to process new images through the already-trained models, significantly reducing computational requirements compared to training models on-the-fly

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system creates a digital copy of the detection function by training machine learning models on image data. Instead of using physical radar sensors to detect objects, the system captures images and creates computational representations (embeddings) of objects through trained neural networks, allowing object detection without the physical radar hardware

Inventive Principle:

Principle #26Copying

Data Source

PatentUS20250131732A1Methods and apparatus to track and classify objects

Publication Date: 2025.04.24 THE BOEING CO

AI summary

Methods and apparatus are disclosed to track and classify objects. An example apparatus for use with an aircraft includes interface circuitry communicatively coupled to an image sensor, the image sensor to capture images, machine readable instructions, and programmable circuitry to at least one of instantiate or execute the machine readable instructions to pre-process the images to identify potential targets, filter, based on the identified potential targets, at least one of the images to determine a presence of a persistent object therein, characterize a looming characteristic of the persistent object, and classify the persistent object based on the looming characteristic meeting or exceeding a looming characteristic threshold.