Hybrid 3D-2D Mapping for Low-Cost Robot Localization
Find Innovative SolutionsGenerate 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
Engineering 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
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
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
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
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
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
3Measurement precision
If three-dimensional sensors are used for environment perception, then measurement precision is improved, but device complexity increases
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
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
Data Source
Figure 1
Figure 2a
Figure 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.