Motion Trajectory Planning With Occluded Obstacle Point Cloud Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion trajectory generation apparatuses for gripping arms fail to account for obstacles behind the gripping target object, leading to potential collisions between the gripping arm and obstacles when the arm is moved according to the generated trajectory.
Innovation Solution
A motion trajectory generation apparatus that uses a depth image sensor to acquire point cloud data, specifies the target object, excludes its point cloud data, estimates and supplements point cloud data for obstacles in the excluded spatial area, and generates a trajectory that avoids interference with both the gripping arm and the target object.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the area of the gripping target object is excluded from the depth image to perform interference determination, then the interference determination can be performed on the target object area, but obstacles behind the target object are not detected and collisions may occur
Solution Approach 1:
The patent performs preliminary actions before excluding the target object area: (1) Acquires depth images from multiple positions including behind the target object, (2) Generates a three-dimensional model of the environment, (3) Identifies obstacles in the excluded area using the three-dimensional model. This preliminary preparation ensures that obstacle information is preserved even when the target object area is excluded for interference determination.
Solution Approach 2:
The patent introduces a three-dimensional model as an intermediary between the depth image and the interference determination. The three-dimensional model serves as a mediator that preserves obstacle information behind the target object while allowing the two-dimensional depth image to be used for interference determination in the target object area.
2Device complexity
If only the second point cloud data (excluding first point cloud data) is used for trajectory planning, then the processing is simpler, but obstacles in the excluded spatial area are not accounted for
Solution Approach 1:
The patent performs preliminary processing to generate a three-dimensional model of the environment and identify obstacles before the trajectory planning stage. This preliminary action ensures that obstacle information is available when needed, without adding complexity to the main trajectory planning process.
Solution Approach 2:
The patent creates a three-dimensional model as a copy or representation of the physical environment. This model contains obstacle information that can be used for trajectory planning without requiring direct processing of the original complex depth image data, thus maintaining simplicity while ensuring reliability.
Data Source
AI summary
An operation processor of the motion trajectory generation apparatus specifies the target object by extracting first point cloud data that corresponds to the target object from a depth image in the vicinity of the target object acquired by a depth image sensor, excludes the first point cloud data from second point cloud data, which is point cloud data in the vicinity of the target object, in the depth image, estimates, using the second point cloud data after the first point cloud data has been excluded, third point cloud data, which is point cloud data that corresponds to an obstacle that is present in a spatial area from which the first point cloud data is excluded in the depth image, and supplements the estimated third point cloud data in the spatial area from which the first point cloud data is excluded, and generates the plan of the motion trajectory.