Hybrid 3D-2D Mapping for Low-Cost Robot Localization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current localization systems for vehicles and robots are either expensive and computationally demanding or limited to simple environments due to the high cost of three-dimensional perception sensors, and inexpensive systems suffer from low robustness due to model assumptions.

Innovation Solution

A method for mapping robust two-dimensional intensity features with their precise three-dimensional positions using a combination of a three-dimensional sensor like LIDAR and a two-dimensional intensity sensor, such as a camera, enabling low-cost localization systems that can accurately navigate complex environments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If expensive three-dimensional sensors like LIDAR are used for accurate localization, then measurement precision is improved, but device cost increases

Engineering Contradiction:
Improvelocalization accuracyVSAvoiddevice cost
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent combines two-dimensional image data from inexpensive cameras with three-dimensional spatial information to create a hybrid mapping system. By integrating 2D intensity features with 3D position data, the system achieves accurate localization without relying solely on expensive three-dimensional sensors, thus resolving the contradiction between measurement precision and device cost

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The invention merges data from multiple sensor types (two-dimensional intensity sensors and three-dimensional position sensors) into a unified mapping system. This combination allows the system to leverage the cost-effectiveness of 2D sensors while incorporating 3D spatial accuracy, thereby achieving high localization precision without the prohibitive cost of exclusively using expensive 3D sensors

Inventive Principle:
Principle #5Merging (Combining)

2Ease of manufacture

If inexpensive localization systems like ultrasonic sensors or mono vision are used, then device cost is reduced, but reliability deteriorates due to model assumptions

Engineering Contradiction:
Improvedevice costVSAvoidrobustness
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent enhances the capabilities of inexpensive two-dimensional sensors by integrating them with three-dimensional position information. This dimensional enhancement allows low-cost systems to overcome their inherent limitations and model assumptions, achieving robust and reliable localization without requiring expensive specialized sensors

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The invention creates a composite mapping system that combines data from different sensor modalities (intensity data and position data). This composite approach allows inexpensive sensor systems to achieve the reliability normally associated with more expensive systems by leveraging multiple data sources to compensate for individual sensor limitations

Inventive Principle:
Principle #40Composite materials

3Measurement precision

If three-dimensional sensors are used for environment perception, then measurement precision is improved, but device complexity increases

Engineering Contradiction:
Improveenvironment perception accuracyVSAvoidsensor system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the environment perception task into two distinct components: two-dimensional intensity feature extraction and three-dimensional position mapping. By dividing the perception function into separate processing streams that are later integrated, the system achieves high measurement precision while managing device complexity through functional segmentation rather than requiring a single complex three-dimensional sensor

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

This approach allows for accurate and cost-effective localization of vehicles and robots in various environments, reducing the need for expensive sensors and improving robustness by associating two-dimensional features with their three-dimensional positions, enabling safe navigation and potential mass production cost reductions.

Implementation Method 1

The precise three-dimensional position may be obtained by a calibrated Light Detection and Ranging (LIDAR) system, for example

Methodology Applied
Scientific EffectLIDAR: LIDAR

Data Source

PatentEP2438401B1Method and apparatus for combining three-dimensional position and two-dimensional intensity mapping for localization
Publication Date: 2019.09.11 ROBERT BOSCH GMBH
  • EP2438401B1 patent drawingFigure 1
  • EP2438401B1 patent drawingFigure 2a
  • EP2438401B1 patent drawingFigure 2b

AI summary

A mapping method includes using a first mobile unit to map two-dimensional features while the first mobile unit traverses a surface. Three-dimensional positions of the features are sensed during the mapping. A three-dimensional map is created including associations between the three-dimensional positions of the features and the map of the two-dimensional features. The three-dimensional map is provided from the first mobile unit to a second mobile unit. The second mobile unit is used to map the two-dimensional features while the second mobile unit traverses the surface. Three-dimensional positions of the two-dimensional features mapped by the second mobile unit are determined within the second mobile unit and by using the three- dimensional map.