Robot Map Merging Using Visual Relative Pose Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing map merging techniques for platoon robots are limited by hardware constraints, requiring initial absolute coordinate information or special sensor equipment to estimate relative poses between robots, and can only merge maps when they have a common overlapping region.
Innovation Solution
A method using an artificial neural network to estimate the relative pose between robots based on images captured by one apparatus, allowing maps created by multiple apparatuses to be merged without relying on special sensors or common overlapping regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional computer-vision feature extraction algorithm is used to merge maps, then map merging can be achieved when common overlapping regions exist, but hardware limitations require initial absolute coordinate information or special sensor equipment to estimate relative pose between robots
Solution Approach 1:
The patent replaces special sensor equipment and mechanical coordinate systems with an artificial neural network-based visual recognition system. The neural network estimates relative pose between robots using only standard RGB camera images, eliminating the need for specialized hardware sensors or pre-configured absolute coordinate information.
Solution Approach 2:
The system uses the robots' own standard RGB cameras to capture images of each other for pose estimation. Each robot serves its own pose estimation needs by capturing images of other robots and processing them through the neural network, making the system self-sufficient without external specialized equipment.
2Adaptability or versatility
If map merging requires common overlapping regions between individual maps, then merging can be performed using traditional algorithms, but the applicability is limited to scenarios where robots have sufficient overlapping coverage
Solution Approach 1:
The patent performs preliminary pose estimation between robots using neural network-based image processing before attempting map merging. By estimating relative poses in advance through visual recognition of robot images, the system can establish proper coordinate transformations even without overlapping map regions, enabling merging in scenarios where traditional methods would fail.
3Measurement precision
If initial absolute coordinate information is provided to set assumptions for pose estimation, then relative pose can be estimated, but the assumption does not match the actual situation and hardware limitations remain
Solution Approach 1:
The patent changes the fundamental parameters used for pose estimation from absolute coordinate information to relative visual features captured by standard RGB cameras. The artificial neural network processes image data to directly estimate relative pose parameters (position and orientation) without requiring pre-set absolute coordinate assumptions, making the system adaptable to actual operational conditions.
Data Source
AI summary
A map merging method for an electronic apparatus which includes: obtaining information about a first local map of a first apparatus, a pose of the first apparatus in the first local map, a second local map of a second apparatus, a pose of the second apparatus in the second local map, and an image of the second apparatus obtained by the first apparatus; identifying a relative pose of the second apparatus relative to the first apparatus from the image using a first trained artificial neural network; transforming the second local map to correspond to the first local map based on the relative pose, the pose of the first apparatus, and the pose of the second apparatus; and merging the first local map and a transformed second local map transformed in the transforming the second local map to output a merged map is provided.