Robot Localization Using Physical Features to Correct SLAM Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current localization and mapping techniques for autonomous robots, such as SLAM, face challenges in maintaining accurate pose estimation and map generation due to drift and error accumulation, particularly in environments with complex layouts and obstacles, which can lead to reduced confidence and inefficient navigation.
Innovation Solution
The method involves maneuvering the robot in both following and coverage modes, using odometry data and sensor feedback to update its pose and map, with re-localization techniques based on physical interactions and template matching to correct errors and maintain confidence, leveraging sensors like encoders, bumpers, and cameras to generate and update maps efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If SLAM techniques are used to build maps and localize robots simultaneously, then autonomous navigation capability is improved, but drift and error accumulation worsen pose estimation accuracy over time
Solution Approach 1:
The patent implements feedback by continuously monitoring robot pose confidence levels and using sensor data from physical interactions with the environment to correct accumulated drift errors. The system feeds back correction information to update the map and pose estimates, maintaining accuracy over time despite continuous operation.
Solution Approach 2:
The robot performs self-correction of pose estimation errors by using its own sensor data from physical interactions (bumpers, encoders, cameras) to detect and correct drift. The system serves itself by autonomously identifying and rectifying its own localization errors without external intervention.
2Measurement precision
If additional sensors are added to improve localization accuracy, then pose estimation precision is improved, but device complexity increases
Solution Approach 1:
The patent makes existing sensors multi-functional by using encoders, bumpers, and cameras for both their primary functions and for pose correction. The encoder data serves both motor control and localization, bumpers serve both obstacle avoidance and feature detection, and cameras serve both environmental perception and pose verification, eliminating the need for dedicated additional sensors.
Solution Approach 2:
The system uses the robot's existing sensor suite to serve dual purposes: primary navigation functions and pose correction. The same sensors that perform basic navigation tasks also provide data for detecting drift and correcting localization accuracy, making the existing sensor system work harder rather than adding new sensors.
3Reliability
If the robot continuously updates pose and map data, then localization accuracy is maintained, but computational resources and processing time are consumed
Solution Approach 1:
The patent applies partial action by updating pose and map data selectively based on confidence level thresholds and trigger events rather than continuously. The system performs full updates only when necessary (when confidence drops below thresholds or specific events occur), reducing computational load while maintaining reliability through targeted corrections.
Solution Approach 2:
The system implements periodic action by scheduling pose corrections at intervals based on confidence level monitoring and operational milestones. Rather than continuous computation, the system periodically evaluates whether updates are needed and performs computations at these discrete intervals, reducing overall energy consumption while maintaining accuracy.
4Speed
If the robot uses odometry data for pose estimation, then navigation speed is improved, but drift causes pose accuracy to deteriorate over distance
Solution Approach 1:
The patent uses feedback to monitor odometry-based pose estimates and correct drift using sensor data from physical interactions with the environment. The system continuously compares expected positions from odometry with actual positions inferred from sensor measurements, feeding back correction signals to eliminate accumulated drift while maintaining navigation speed.
Solution Approach 2:
The patent introduces sensor data from physical interactions as an intermediary to mediate between odometry estimates and actual pose. This intermediary data serves as a reference point to detect and correct drift, allowing the system to maintain fast odometry-based navigation while periodically correcting accuracy through environmental feature detection.
Data Source
AI summary
A method includes maneuvering a robot in (i) a following mode in which the robot is controlled to travel along a path segment adjacent an obstacle, while recording data indicative of the path segment, and (ii) in a coverage mode in which the robot is controlled to traverse an area. The method includes generating data indicative of a layout of the area, updating data indicative of a calculated robot pose based at least on odometry, and calculating a pose confidence level. The method includes, in response to the confidence level being below a confidence limit, maneuvering the robot to a suspected location of the path segment, based on the calculated robot pose and the data indicative of the layout and, in response to detecting the path segment within a distance from the suspected location, updating the data indicative of the calculated pose and/or the layout.