Robotic Obstacle Detection via Covariance Eigenvectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional obstacle detection techniques for robotic devices using Lidar or time-of-flight sensors are slow due to reliance on k-dimensional (KD) tree graphs, consuming excessive computing resources and delaying obstacle identification, which can lead to unsafe environments and property damage.
Innovation Solution
A computing device processes time-of-flight sensor data to generate shifted depth images, calculate Euclidean distances, and determine eigenvectors for covariance matrices without constructing a KD tree graph, segmenting floor points and clustering obstacles to expedite obstacle detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional obstacle detection techniques using KD tree graphs are employed, then obstacle detection accuracy is maintained, but processing speed deteriorates and computing resources are excessively consumed
Solution Approach 1:
The patent segments the point cloud data processing into distinct stages: generating candidate points from depth images, filtering candidates using geometric constraints, and performing final obstacle detection. This segmentation eliminates the need for comprehensive KD tree construction while maintaining detection accuracy through targeted processing of suspicious regions only.
Solution Approach 2:
The patent extracts only the necessary computational elements for obstacle detection by directly calculating Euclidean distances from depth image data and identifying candidate points based on geometric constraints, rather than extracting and organizing all points into a KD tree structure. This extraction approach retrieves only the critical information needed for detection.
2Measurement precision
If conventional obstacle detection techniques using KD tree graphs are employed, then comprehensive obstacle analysis is achieved, but computing resources are excessively consumed
Solution Approach 1:
The processing pipeline is segmented to handle only suspicious regions identified from depth image gradients and geometric constraints, rather than processing the entire point cloud. This segmentation reduces computing resource consumption by focusing computational effort only where obstacles are likely to be present.
Solution Approach 2:
The patent applies partial action by performing detailed obstacle analysis only on candidate points that satisfy geometric constraints and are located in suspicious regions, rather than analyzing all points. This partial processing approach maintains detection completeness while significantly reducing computing resource consumption.
3Measurement precision
If conventional obstacle detection techniques are employed, then thorough obstacle identification is achieved, but time consumption increases
Solution Approach 1:
The patent performs preliminary actions by generating candidate points and identifying suspicious regions using depth image gradients and geometric constraints before conducting final obstacle detection. This preliminary filtering reduces the dataset size early in the process, maintaining identification accuracy while reducing overall detection time.
Solution Approach 2:
The patent skips the time-consuming KD tree construction and nearest neighbor search steps by directly calculating distances from depth images and using geometric constraints to identify candidate points. This skipping of unnecessary intermediate steps maintains thorough obstacle identification while significantly reducing detection time.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach accelerates obstacle detection, conserving computing resources and enabling efficient navigation by identifying edges and generating navigation instructions, thus preventing collisions and property damage.
Implementation Method 1
A robotic device may utilize Lidar or time-of-flight sensor data to detect the obstacles in the environment
Data Source
AI summary
A device may receive, from a robotic device, time-of-flight sensor data that includes a point cloud derived from depth images of an environment with a floor and one or more obstacles, and may shift the depth images by a number of pixels to generate shifted depth images. The device may subtract the shifted depth images from the depth images to generate final images, and may calculate Euclidean distances associated with the final images. The device may generate masks for the final images associated with Euclidean distances that are less than a search radius, and may calculate a covariance matrix based on the masks. The device may calculate eigenvectors for the covariance matrix, and may generate an occupancy grid for the environment based on the eigenvectors for the covariance matrix.