LIDAR Recalibration via Point Cloud Offset Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Autonomous driving vehicles face challenges in maintaining accurate sensor calibration, particularly for LIDAR systems, which is crucial for perception and prediction, as existing recalibration methods are not efficient or timely, leading to potential navigation errors.

Innovation Solution

A point clouds-based light detection ranging (LIDAR) recalibration system that extracts feature information from stored three-dimensional maps, identifies matching features, calculates average offset distances, and sends alerts for recalibration when predetermined conditions are met, ensuring accurate sensor alignment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional recalibration methods are used, then sensor alignment may be maintained, but recalibration is not timely and navigation errors occur

Engineering Contradiction:

Improvesensor alignment accuracyVSAvoidrecalibration timing

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system continuously monitors the alignment status of LIDAR sensors by comparing point cloud data with pre-stored map data, and provides feedback when misalignment exceeds thresholds, triggering timely recalibration alerts to maintain navigation accuracy

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary calibration by storing accurate map data and establishing baseline alignment parameters before actual navigation occurs, enabling proactive detection and correction of drift before significant errors accumulate

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If continuous monitoring is implemented, then navigation accuracy is improved, but computational resources and processing time increase

Engineering Contradiction:

Improvenavigation accuracyVSAvoidcomputational energy consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system performs partial monitoring by selectively comparing point cloud data against pre-stored map features rather than processing all data continuously, and uses threshold-based triggering to activate full recalibration procedures only when necessary, reducing computational overhead while maintaining accuracy

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system pre-processes and stores reference map data and alignment parameters before navigation, enabling real-time monitoring to focus on detecting deviations rather than performing complete spatial matching, significantly reducing computational energy requirements during operation

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12123984B2Point clouds based lidar recalibration system for autonomous vehicles

Publication Date: 2024.10.22 BAIDU USA LLC

AI summary

Embodiments of the present disclosures disclose a method and a system to notify an operator that perception sensors of an autonomous driving vehicle (ADV) need to be recalibrated. In one embodiment, a system perceives a surrounding environment of an autonomous driving vehicle (ADV), including one or more obstacles. The system extracts feature information from previously stored point clouds mapping a three-dimensional surrounding environment of the ADV. The system identifies one or more matching features between the extracted feature information and features of the one or more obstacles. The system determines an average offset distance based on each of the matching features. The system determines an average offset distance distribution based on the average offset distance over a period of time. The system sends an alert to the ADV to alert that the one or more sensors is recommended for recalibration if the average offset distance distribution satisfies a predetermined condition.