SLAM Graph Node Pruning to Preserve Accuracy on Mobile Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing size of SLAM graphs in mobile devices due to constant node and edge additions leads to higher computing power and time requirements, but thinning these graphs to reduce data volume often compromises map accuracy.
Innovation Solution
A method that removes nodes from the SLAM graph based on scale-invariant density, prioritizing nodes with high density around them, and re-links edges to maintain accuracy, using geometric methods to minimize information loss and avoid false loop-closure edges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If nodes and edges are constantly added to the SLAM graph to improve mapping completeness, then the map accuracy and environmental coverage are improved, but the computing power requirements and processing time increase
Solution Approach 1:
The patent extracts and removes specific nodes from the SLAM graph based on scale-invariant density analysis. By identifying and removing nodes with high density (redundant measurements), the system reduces graph size and computing requirements while preserving essential mapping information. This extraction process resolves the contradiction by eliminating unnecessary data that contributes to computational load without compromising map accuracy.
Solution Approach 2:
The patent changes the parameter of node selection by using scale-invariant density as the criterion for node removal. This parameter transformation allows the system to identify redundant nodes objectively, enabling graph thinning that maintains mapping precision while reducing computational complexity. The scale-invariant property ensures the method works across different map scales.
2Productivity
If the SLAM graph is thinned to reduce data volume, then the processing speed and memory usage are improved, but the map accuracy deteriorates
Solution Approach 1:
The patent applies local quality by differentiating between important and redundant nodes based on their local density characteristics. Instead of uniform node removal, the system selectively removes nodes in high-density regions while preserving nodes in low-density regions that contain critical mapping information. This localized approach maintains map accuracy in essential areas while enabling graph thinning in redundant areas, thus improving processing speed without sacrificing overall map precision.
Solution Approach 2:
The patent removes only the necessary portion of nodes (those with high scale-invariant density) rather than uniformly thinning the entire graph. This partial action approach ensures that only redundant information is discarded while essential mapping data is retained, thereby improving processing speed while maintaining sufficient map accuracy for navigation purposes.
3Quantity of substance
If arbitrary nodes are removed from the SLAM graph to reduce size, then the data volume is reduced, but false loop-closure edges may be created and map reliability decreases
Solution Approach 1:
The patent performs preliminary analysis of node density and connectivity before removing any nodes. By calculating scale-invariant density metrics in advance, the system identifies safe candidates for removal that are unlikely to create false loop-closures. This preliminary assessment ensures that node removal maintains graph integrity and prevents reliability degradation while achieving data volume reduction.
Solution Approach 2:
The patent incorporates feedback mechanisms to monitor the impact of node removal on graph structure and map reliability. By continuously assessing whether removed nodes were critical for loop-closure detection, the system can adjust its thinning strategy to prevent false loop-closures. This feedback-driven approach ensures that data volume reduction does not compromise map reliability.
Data Source
AI summary
A method for thinning a SLAM graph, which is used for operating a mobile device and has a multiplicity of nodes and a multiplicity of edges, each of which ends with an end point at a node. The SLAM graph is obtained, nodes are removed from the SLAM graph, and then an updated SLAM graph is output. Removing a node in each case includes determining, out of the multiplicity of nodes, the node whose scale-invariant density at further nodes around said node is the highest, and removing the determined node whose density is the highest from the SLAM graph.