Grid-Based Point Cloud Clustering for Autonomous Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Processing point cloud data is computationally intensive and challenging for mobile platforms, particularly in dynamic environments, leading to reduced situational awareness and navigation difficulties due to the complexity of handling large volumes of data.
Innovation Solution
A system that overlays a grid onto a point cloud to group points into discrete cells, computes feature vectors for these cells, and uses a machine learning clustering model to identify and cluster similar entities, reducing the need for individual point analysis and improving processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If point cloud data is processed individually at the point level, then clustering accuracy is improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent divides the point cloud data into discrete grid cells, transforming the continuous space into discrete units. This segmentation allows the system to process groups of points collectively rather than individually, reducing computational complexity while maintaining clustering accuracy through the preservation of spatial relationships within each cell
Solution Approach 2:
The patent introduces a grid cell dimension as an intermediate level of abstraction between individual points and clusters. By operating at the grid cell level rather than the point level, the system adds a dimensional layer that reduces the number of entities to process while retaining sufficient spatial information for accurate clustering
2Loss of information
If all points in the point cloud are processed, then complete environmental awareness is achieved, but processing time increases
Solution Approach 1:
The point cloud is segmented into grid cells that can be processed independently and in parallel. This segmentation enables the system to process spatially distributed data units simultaneously, reducing overall processing time while ensuring complete environmental coverage through the union of all processed cells
Solution Approach 2:
The patent processes grid cells selectively based on their relevance to the current task rather than processing every single point uniformly. By applying partial action to the most significant regions and using the grid structure to organize processing priorities, the system achieves complete environmental awareness with reduced processing time
Data Source
AI summary
System, methods, and other embodiments described herein relate to improving clustering of points within a point cloud. In one embodiment, a method includes grouping the points into cells of a grid. The grid divides an observed region of a surrounding environment associated with the point cloud into the cells. The method includes computing feature vectors for the cells that use cell features to characterize the points in the cells and relationships between the cells. The method includes analyzing the feature vectors according to a clustering model to identify clusters for the cells. The clustering model evaluates the cells to identify which of the cells belong to common entities. The method includes providing the clusters as assignments of the points to the entities depicted in the point cloud.