Mobile Robot Path Control During SLAM Localization Failure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for mobile robots to recover from localization failures, such as those encountered in SLAM systems, are inefficient and computationally expensive, especially in environments with repetitive patterns or dynamic obstacles, leading to low success rates and slowed navigation.
Innovation Solution
A method involving a monovision SLAM system that splits images into vertical portions to identify passable paths, allowing the robot to continue moving forward and avoid obstacles without requiring expensive computational resources, using image segmentation techniques like Mean Shift to differentiate between floor and background.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the robot stops and spins to re-localize using sensor data, then the robot attempts to recover from localization failure, but the probability of success is low and navigation speed significantly decreases
Solution Approach 1:
The system performs preliminary localization attempts using the current forward-facing image before committing to the spin maneuver. By evaluating the current image for localization feasibility first, the system avoids unnecessary spins and maintains navigation speed while improving recovery success rate through preliminary action.
2Reliability
If three-dimensional images are constructed from two-dimensional image data using deep neural networks, then the robot can travel safely avoiding obstacles, but expensive computational resources are required and the process is time-consuming
Solution Approach 1:
The system extracts only the essential depth information needed for obstacle avoidance by analyzing vertical portions of the image, rather than constructing complete three-dimensional images. This extraction approach provides sufficient obstacle avoidance capability while dramatically reducing computational complexity and processing time.
3Reliability
If heuristic methods are used to reconstruct the three-dimensional scene, then the robot can navigate static environments with precise geometric properties, but the method struggles with dynamic obstacles such as human beings
Solution Approach 1:
The system dynamically adapts its localization and navigation approach by continuously evaluating current images and adjusting the passable path selection in real-time. This dynamic approach allows the robot to handle both static environments with precise geometric properties and dynamic obstacles like human beings, improving versatility while maintaining navigation accuracy.
Data Source
AI summary
Described is a method of controlling movement of a mobile robot in the event of a localization failure. The method comprises, on an occurrence of the localization failure, receiving an image from an image sensor of the mobile robot. The method includes splitting received image into at least two vertical portions and determining if one of said at least two vertical portions is indicative of a passable path. The method includes controlling the mobile robot to travel along the passable path indicated by a selected one of said at least two vertical portions.