Vehicle LiDAR Object Separation for Stable Autonomous Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing vehicle control systems using LiDAR often inaccurately identify external objects, particularly structured and unstructured objects, which can compromise driving stability in assist and autonomous modes.

Innovation Solution

A vehicle control apparatus and method that utilizes LiDAR data to determine virtual boxes corresponding to external objects, applies a neural network model to analyze LiDAR points, and separates structured and unstructured objects using a Gaussian mixture model (GMM) to improve object classification and driving stability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Difficulty of detecting and measuring

If LiDAR is used to identify external objects, then object detection capability is improved, but object identification accuracy deteriorates due to incorrect identification of structured and unstructured objects

Engineering Contradiction:

Improveobject detection capabilityVSAvoidobject identification accuracy

Core Design Contradiction:

Difficulty of detecting and measuringVSMeasurement precision

Solution Approach 1:

The patent segments the detection space into multiple virtual boxes, each dedicated to detecting specific types of objects (structured objects like vehicles and unstructured objects like pedestrians). This segmentation allows the system to apply type-specific detection algorithms to each virtual box, improving overall identification accuracy while maintaining comprehensive detection coverage

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different detection qualities and algorithms to different virtual boxes based on their intended object types. Structured object virtual boxes use algorithms optimized for vehicle detection, while unstructured object virtual boxes use algorithms optimized for pedestrian detection. This local quality approach resolves the contradiction by making the detection system adapt its precision to the specific object type in each region

Inventive Principle:

Principle #3Local quality

2Area of stationary object

If virtual boxes are used to detect external objects, then detection coverage is improved, but object classification accuracy deteriorates due to inability to distinguish structured and unstructured objects

Engineering Contradiction:

Improvedetection coverageVSAvoidobject classification accuracy

Core Design Contradiction:

Area of stationary objectVSMeasurement precision

Solution Approach 1:

The detection space is divided into multiple virtual boxes, each assigned to detect specific object types. This segmentation enables the system to maintain broad detection coverage across all object types while simultaneously achieving high classification accuracy within each virtual box through specialized detection algorithms

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces virtual boxes as intermediary detection zones between the LiDAR sensor and the object classification process. Each virtual box acts as a mediator that pre-processes and categorizes objects by type before final classification, thereby improving both coverage and accuracy

Inventive Principle:

Principle #24Intermediary (Mediator)

3Adaptability or versatility

If combination virtual box is used to detect multiple objects, then detection versatility is improved, but driving stability deteriorates due to incorrect separation of structured and unstructured objects

Engineering Contradiction:

Improvedetection versatilityVSAvoiddriving stability

Core Design Contradiction:

Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The combination virtual box is segmented into multiple sub-regions, each dedicated to detecting specific object types (structured objects like vehicles and unstructured objects like pedestrians). This segmentation enables the system to maintain versatility in detecting multiple object types simultaneously while ensuring driving stability through accurate separation and classification of each object type using specialized algorithms

Inventive Principle:

Principle #1Segmentation

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Enhances the accuracy of object identification, enabling better vehicle control and stability in driving assist and autonomous modes by distinguishing between structured and unstructured objects.

Implementation Method 1

light detection and ranging (LiDAR) device... The LiDAR device may be configured to obtain sensing information corresponding to a first external object

Methodology Applied

Scientific EffectLight detection and ranging (LiDAR): LIDAR

Data Source

PatentUS20250349127A1Vehicle control apparatus and method thereof

Publication Date: 2025.11.13 HYUNDAI MOTOR CO LTD

AI summary

A vehicle control apparatus and a method thereof are provided. The vehicle control apparatus includes light detection and ranging (LiDAR) device and a processor. The LiDAR device is configured to obtain sensing information corresponding to a first external object, and a processor. The processor is configured to determine, based on the sensing information, a first virtual box, determine a candidate group including a combination virtual box. The combination virtual box includes the first virtual box and a second virtual box. The processor is further configured to determine, based on applying the LiDAR data to a neural network model, a distribution of the LiDAR points, divide, based on the distribution, the combination virtual box into an adjusted first virtual box and an adjusted second virtual box, and control, based on at least one of the adjusted first virtual box or the adjusted second virtual box, an operation of a vehicle.