Automated Lidar Map Alignment Using Feature Matching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Manual alignment of LIDAR maps is inaccurate and labor-intensive due to positioning errors and environmental changes, requiring frequent realignment of maps in vehicle fleets with varying sensor heights and environmental modifications.

Innovation Solution

Automated alignment method that converts LIDAR maps into feature sets, calculates a transformation using matcher algorithms, and aligns subsequent maps based on a manually aligned initial map, reducing manual intervention and improving accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual alignment of LIDAR maps is performed using a user interface, then the operator can rotate or pan the map, but the alignment accuracy deteriorates due to positioning errors and environmental changes

Engineering Contradiction:
Improvemanual alignment operationVSAvoidalignment accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The system performs automatic alignment using LIDAR data matching and transformation algorithms, eliminating the need for manual operator intervention. The electronic control unit automatically calculates transformation parameters and aligns maps based on feature point matching between LIDAR scans and existing maps, making the system self-aligning without human input.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The manual mechanical alignment process using GUI rotation and panning is replaced by an automated computational system. The electronic control unit uses LIDAR point cloud data, feature extraction algorithms, and mathematical transformation (rotation matrices, translation vectors) to achieve precise alignment, substituting manual mechanical operations with automated computational geometry.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Manufacturing precision

If manual post-processing alignment is performed for each map, then the map can be aligned to semantic positions, but the time and labor required increase significantly

Engineering Contradiction:
Improvemap alignment precisionVSAvoidalignment time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs preliminary alignment by automatically calculating transformation parameters during the mapping process itself. When a new map is created, the electronic control unit immediately computes the transformation needed to align it with the existing map using LIDAR feature matching, so that alignment is accomplished before any manual intervention would be required, eliminating subsequent post-processing time.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The alignment process becomes continuous and automated rather than discrete and manual. As each new LIDAR map is generated, the system continuously performs automatic alignment through feature point matching and transformation calculation, maintaining an always-aligned state without interrupting the mapping workflow for manual alignment operations.

Inventive Principle:
Principle #20Continuity of useful action

3Adaptability or versatility

If LIDAR sensors are installed at different heights in vehicles, then fleet deployment flexibility increases, but map alignment consistency deteriorates

Engineering Contradiction:
Improvefleet deployment flexibilityVSAvoidmap alignment consistency
Core Design Contradiction:
Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The system automatically adapts to different sensor heights by calculating unique transformation parameters for each vehicle's LIDAR data. The electronic control unit determines the specific vertical offset and rotational transformation needed for each sensor position, adjusting the alignment parameters dynamically based on the actual sensor installation characteristics of each fleet vehicle.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The alignment system becomes universal and works with LIDAR sensors installed at any height or position. By using feature-based LIDAR matching rather than height-dependent assumptions, the system handles diverse sensor configurations across the fleet through a single automated process, making the alignment solution universally applicable to all vehicles regardless of sensor installation variations.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3514493B1Method for aligning cards of a lidar system
Publication Date: 2024.07.10 ROBERT BOSCH GMBH
  • EP3514493B1 patent drawingFigure 1~2

AI summary

The present invention relates to a method for aligning maps of an electronic positioning system, in particular a LIDAR positioning system used in a driver assistance system of a motor vehicle or industrial truck, in automated driving of motor vehicles, industrial trucks or mobile robots, in which it is particularly provided that at least one map (215', 305) of the electronic positioning system to be aligned is aligned (240, 320) with respect to a building model (220) and/or with respect to an already aligned map (300). When aligning (240, 320) the at least one map (215', 305) to be aligned, a different installation height of at least one positioning sensor can be taken into account.