Frontier Goal Selection for Occlusion-Aware Autonomous Map Building
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic devices face inefficiencies in autonomous map building due to reliance on Euclidean distances and high-resolution navigational maps, which do not consider occlusions or overall environmental factors, leading to suboptimal path planning and increased computational costs.
Innovation Solution
Implementing a method that determines trajectory costs, co-visibility costs, and frontier costs for robotic devices to select the most efficient frontier goals by combining factors such as path friendliness, contextual information, and co-visibility analysis, using lower resolution maps for goal selection and higher resolution maps for navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Euclidean distances and high-resolution navigational maps are used for path planning, then navigation precision is improved, but computational costs increase and environmental factors such as occlusions are not considered
Solution Approach 1:
The patent segments the map into two resolutions: high-resolution navigational maps for precise navigation and lower-resolution frontier cost maps for goal selection. This segmentation allows the system to use computationally intensive high-resolution data only when necessary for navigation, while using lighter lower-resolution data for the more frequent frontier goal selection process, thereby reducing overall computational costs while maintaining navigation precision.
Solution Approach 2:
The patent applies local quality by using different map resolutions for different functional purposes. High-resolution maps are used locally for navigation tasks requiring precision, while lower-resolution maps are used for frontier goal selection where absolute precision is less critical. This differentiated approach optimizes computational resource allocation based on the specific requirements of each subsystem.
2Measurement precision
If high-resolution navigational maps are used for frontier goal selection, then positioning accuracy is improved, but computational efficiency deteriorates
Solution Approach 1:
The patent divides the mapping and goal selection process into two tiers: high-resolution navigational maps for accurate positioning and lower-resolution frontier cost maps for efficient goal evaluation. By segmenting the computational workload this way, the system achieves positioning accuracy when needed while maintaining computational efficiency during the more frequent frontier goal selection operations.
Solution Approach 2:
The patent applies partial action by using lower-resolution maps for frontier goal selection, which provides sufficient information for comparing relative costs without the full detail of high-resolution maps. This partial use of resolution optimizes the balance between positioning accuracy and computational efficiency by applying only the necessary level of detail for each specific task.
3Ease of manufacture
If traditional path planning methods are used, then implementation simplicity is maintained, but map building efficiency deteriorates due to not considering occlusions and environmental factors
Solution Approach 1:
The patent merges multiple factors including trajectory costs, co-visibility costs, and frontier costs into a unified frontier cost calculation. This combination integrates considerations of occlusions and environmental factors that were previously handled separately or ignored, thereby improving map building efficiency while maintaining a relatively simple implementation through the unified cost function approach.
Solution Approach 2:
The patent introduces co-visibility cost as an intermediary factor that mediates between the robot's current position and frontier goals by accounting for occlusions and environmental factors. This intermediary calculation allows the system to efficiently evaluate which frontier goals are actually accessible and worthwhile to explore, improving map building efficiency without significantly complicating the overall system architecture.
Data Source
AI summary
Embodiments include methods executed by a processor of a robotic device for selecting a frontier goal for autonomous map building within a space, including determining trajectory costs for the robotic device traversing from an occupied area to each of a plurality of available frontier goals, wherein the plurality of available frontier goals include positions within the space from which the robotic device is configured to collect information to autonomously build a map of the space, determining co-visibility costs for each of the plurality of available frontier goals, determining a frontier cost for each of the plurality of available frontier goals based on a difference between the trajectory cost and the co-visibility cost of the respective available frontier goal; and selecting one of the plurality of available frontier goals having a lowest determined frontier cost.