SLAM Landmark Association Using Representative Measurements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current SLAM methods require significant computational resources to determine whether all acquired measurements are associated with pre-recognized landmarks, hindering efficiency and increasing the likelihood of missing new landmarks.
Innovation Solution
The method involves selecting representative measurements through clustering or grid cells and determining their association with pre-recognized landmarks, reducing computational load and enhancing the probability of identifying new landmarks by selecting a preset number of landmarks based on Euclidean distance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all acquired measurements are checked for association with pre-recognized landmarks, then measurement precision is improved, but computational resources are excessively consumed
Solution Approach 1:
The patent segments the set of all measurements into a subset of representative measurements using clustering algorithms (e.g., selecting cluster centers) or grid cell methods. This segmentation allows the system to process only the essential measurements for landmark identification, thereby reducing computational resource consumption while maintaining identification accuracy.
Solution Approach 2:
The patent applies partial action by checking association with only a preset number of pre-recognized landmarks (e.g., nearest neighbors) rather than all available landmarks. This partial checking approach reduces computational load while still ensuring reliable landmark identification through selective verification.
2Measurement precision
If all acquired measurements are checked for association with pre-recognized landmarks, then measurement precision is improved, but processing speed deteriorates
Solution Approach 1:
The patent segments measurements into representative subsets through clustering or grid cell approaches, reducing the number of measurements that require detailed processing. This segmentation maintains identification accuracy while significantly improving processing speed by focusing computational effort on essential measurements only.
Solution Approach 2:
The patent implements partial action by limiting landmark association checks to a preset number of candidate landmarks (e.g., nearest neighbors) rather than exhaustively checking all landmarks. This approach accelerates processing while preserving identification precision through selective verification of the most promising candidates.
3Productivity
If representative measurements are selected through clustering or grid cells, then productivity is improved, but device complexity increases
Solution Approach 1:
The patent employs segmentation through well-established clustering algorithms (e.g., K-means) or grid cell partitioning methods. These are standard techniques in data processing that organize measurements into manageable groups, improving SLAM processing efficiency while adding only moderate algorithmic complexity that is widely understood and implementable.
Data Source
AI summary
The present disclosure relates to a method of selecting representative measurements from among measurements related to an environment around a movable object and determining whether the representative measurements are associated with landmarks to realize SLAM and an electronic apparatus for the same. One or more of an electronic apparatus, a vehicle, and an autonomous vehicle of the present disclosure may be connected to, for example, an artificial intelligence module, an unmanned aerial vehicle (UAV), a robot, an augmented reality (AR) device, a virtual reality (VR) device, or a 5G service device.