LiDAR Landmark Shape Fitting for Precise Autonomous Localization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional perception and localization methods for autonomous vehicles using LiDAR data are limited in functionality, accuracy, and precision, especially when detecting lane lines, which hinders effective navigation.

Innovation Solution

A deep neural network (DNN) is used to generate per-point classifications of measured 3D points and regress the shape of detected landmarks, such as lane lines, using connected components analysis and shape fitting techniques to improve localization precision.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If per-point classification of LiDAR data using deep neural networks is used, then localization capability is enabled, but localization precision is insufficient due to sparse LiDAR points

Engineering Contradiction:

Improvelocalization capabilityVSAvoidlocalization precision

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent combines multiple LiDAR scans over time to accumulate point cloud data, merging sparse individual scans into a denser composite representation that improves localization precision while maintaining the ability to function with LiDAR-only input

Inventive Principle:

Principle #5Merging (Combining)

2Ease of manufacture

If Hough transform is used to find lines in classifications, then straight lines can be detected, but it becomes inefficient and ineffective for curved roads

Engineering Contradiction:

Improvedetection efficiencyVSAvoidhandling curved roads

Core Design Contradiction:

Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent replaces the Hough transform's straight-line assumption with curve-fitting algorithms that can model curved lane lines and road boundaries, enabling effective detection on curved roads while maintaining computational efficiency through optimized fitting procedures

Inventive Principle:

Principle #14Spheroidality (Curvature)

3Device complexity

If conventional perception methods are used, then system complexity is reduced, but functionality and accuracy are limited

Engineering Contradiction:

Improvesystem simplicityVSAvoidlocalization accuracy

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent replaces conventional geometric and threshold-based perception methods with deep neural network-based classification and regression systems, achieving superior localization accuracy and functionality while managing complexity through end-to-end learning approaches

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS12482118B2Landmark perception for localization in autonomous systems and applications

Publication Date: 2025.11.25 NVIDIA CORP

AI summary

In various examples, perception of landmark shapes may be used for localization in autonomous systems and applications. In some embodiments, a deep neural network (DNN) is used to generate (e.g., per-point) classifications of measured 3D points (e.g., classified LiDAR points), and a representation of the shape of one or more detected landmarks is regressed from the classifications. For each of one or more classes, the classification data may be thresholded to generate a binary mask and/or dilated to generate a densified representation, and the resulting (e.g., dilated, binary) mask may be clustered into connected components that are iteratively: fitted a shape (e.g., a polynomial or Bezier spline for lane lines, a circle for top-down representations of poles or traffic lights), weighted, and merged. As such, the resulting connected components and their fitted shapes may be used to represent detected landmarks and used for localization, navigation, and/or other uses.