Mobile Robot Mapping With Particle-Graph Trajectory Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large-scale scenes, mobile robots face significant trajectory deviations due to particle filtering and graph optimization methods failing to accurately match and converge, leading to errors in map construction.
Innovation Solution
A mapping method for mobile robots that combines particle filtering and graph optimization, where particle positions are estimated using odometry data, matched with obstacle information, and weights are normalized to ensure even distribution, followed by resampling and graph optimization to correct trajectories, resulting in a more accurate map construction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If particle filtering approach is used to locate the robot in large-scale scenes, then the robot can obtain position estimates, but the particle group cannot match and converge to the correct position, leading to large trajectory deviation
Solution Approach 1:
The patent combines particle filtering approach with graph optimization approach to create a hybrid localization system. The particle filter provides initial position estimates while graph optimization corrects trajectory deviations by optimizing the consistency of loop closures and path constraints, thereby resolving the convergence issue of particle groups and improving overall trajectory accuracy.
Solution Approach 2:
The system implements feedback mechanisms where the graph optimization module continuously receives trajectory data from particle filtering and provides correction feedback. The optimized trajectory information is fed back to adjust particle weights and positions, creating a closed-loop system that progressively reduces trajectory deviations and improves localization reliability.
2Productivity
If graph optimization approach is used to locate the robot, then the robot can optimize its path, but measurement errors are generated during single path updates, leading to large trajectory deviation
Solution Approach 1:
The patent integrates graph optimization with particle filtering to create a mutually compensating system. While graph optimization efficiently processes path updates, the particle filtering component provides probabilistic position estimates that help correct measurement errors. The combination allows the system to maintain high path optimization efficiency while improving trajectory accuracy through the complementary strengths of both methods.
3Productivity
If traditional mapping methods are used in large-scale scenes, then the robot can construct a map, but trajectory deviations cause errors in the constructed map
Solution Approach 1:
The system employs feedback loops where the graph optimization module continuously refines the robot's trajectory estimates and feeds this corrected information back to the map building process. This ensures that the constructed map is based on accurate trajectory data, eliminating the propagation of trajectory deviations into mapping errors and significantly improving map precision in large-scale environments.
Data Source
AI summary
A mapping method for a mobile robot includes: obtaining positions of particles corresponding to the mobile robot using a particle filtering approach; selecting one of the particles with a largest weight in a particle swarm corresponding to a current time point, performing position matching between the one of the particles and a historical trajectory formed by the one of the particles with the largest weight, and determining whether the positions of the particles corresponding to the mobile robot match positions of historical trajectory points of the mobile robot; in response to the positions of the particles corresponding to the mobile robot matching the positions of the historical trajectory points of the mobile robot, optimizing a trajectory of the mobile robot using a graph optimization approach; and building a map based on the optimized trajectory.