Vision-Based Voxel Occupancy Modeling for Autonomous Navigation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing autonomous navigation systems struggle to accurately predict the occupancy status of surrounding environments, which is crucial for safe and reliable operation of autonomous vehicles and robots.

Innovation Solution

A trained artificial intelligence model processes image data from cameras to predict the occupancy status of voxels in the surroundings, generating a dataset and graphical indicators for autonomous decision-making.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional sensor-based methods are used for occupancy detection, then the system can detect objects in the environment, but the measurement precision and reliability of occupancy prediction are insufficient

Engineering Contradiction:

Improveoccupancy prediction accuracyVSAvoidsystem reliability

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent replaces traditional mechanical sensor-based occupancy detection systems with an artificial intelligence model that processes image data. This substitution enables more accurate occupancy prediction by using deep learning algorithms to analyze visual information and determine voxel occupancy status, thereby improving measurement precision while maintaining system reliability through software-based probabilistic reasoning.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If AI models are used to predict occupancy status, then prediction accuracy improves, but computational complexity and processing time increase

Engineering Contradiction:

Improveoccupancy prediction accuracyVSAvoidcomputational complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the three-dimensional space into discrete voxels and processes occupancy prediction for each voxel independently through the AI model. This segmentation approach allows the system to handle complex computational tasks in manageable units, improving prediction accuracy at the voxel level while reducing overall computational complexity through parallel processing capabilities.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements a two-stage processing approach where the AI model first performs featurization of image data, then applies the trained model for occupancy prediction. This partial action strategy processes only the most relevant features extracted from images, reducing unnecessary computational overhead while maintaining high prediction accuracy for critical occupancy determination.

Inventive Principle:

Principle #16Partial or excessive action

3Area of stationary object

If multiple camera feeds are processed for comprehensive environmental understanding, then occupancy detection coverage improves, but data processing time and computational load increase

Engineering Contradiction:

Improveenvironmental coverageVSAvoidprocessing time

Core Design Contradiction:

Area of stationary objectVSLoss of time

Solution Approach 1:

The patent merges multiple camera feeds into a unified three-dimensional space representation before processing. By combining data from multiple cameras and projecting it onto a common voxel grid, the system achieves comprehensive environmental coverage while reducing processing time through consolidated data structures and shared computational pathways across all camera inputs.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent performs preliminary temporal alignment and synchronization of multiple camera feeds before they are input to the AI model. This preliminary action ensures that data from all cameras corresponds to the same time moment, improving occupancy detection accuracy while reducing processing time by eliminating the need for post-processing temporal corrections and enabling more efficient batch processing.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12469160B2Artificial intelligence modeling techniques for vision-based occupancy determination

Publication Date: 2025.11.11 TESLA INC

AI summary

Disclosed herein are methods and systems for using artificial intelligence modeling techniques to train and execute an artificial intelligence model to analyze camera feed received from an ego to generate an occupancy data indicating whether different voxels within the ego's surroundings are occupied by an object having mass. A method comprises inputting, using a camera of an ego object, image data of a space around the ego object into an artificial intelligence model; predicting, by executing the artificial intelligence model, an occupancy attribute of a plurality of voxels; and generating a dataset based on the plurality of voxels and their corresponding occupancy attribute.