Trajectory Generation Device Using DBiRRT for Robot Arm Path Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing trajectory generation methods for robot arms are inefficient when the generated trajectory does not correspond to any known pattern, leading to suboptimal performance in complex and changing environments.
Innovation Solution
A trajectory generation device that includes a storage unit for multiple trajectories, a trajectory acquisition unit to select a similar environment trajectory, and a trajectory generation unit using the Differential Bi-Directional Rapidly-Exploring Random Tree (DBiRRT) method to calculate and connect the longest trajectory part, excluding obstacles, and smooth the trajectory to connect both ends to predetermined start and end points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a trajectory generation method selects from registered states in a known-state registration tree, then it can generate trajectories for known patterns, but it cannot efficiently utilize registered information when the trajectory does not correspond to any known pattern
Solution Approach 1:
The system pre-registers multiple trajectories in advance in the storage unit, creating a database of known trajectories. When generating a new trajectory, the system retrieves and adapts from these pre-registered trajectories rather than calculating from scratch, improving efficiency for known patterns while maintaining adaptability through selective retrieval and modification.
Solution Approach 2:
The trajectory acquisition unit copies trajectories from the storage unit that correspond to environments similar to the current environment. Instead of generating entirely new trajectories, the system replicates and adapts existing trajectories from the database, efficiently leveraging registered information while adjusting to current conditions.
2Reliability
If the system calculates a complete trajectory from start to end points, then it ensures full path coverage, but it increases computational complexity and time consumption
Solution Approach 1:
The trajectory is divided into segments: a longest trajectory part extracted from the acquired trajectory, and connection trajectories generated by RRT method from start point and end point to the ends of the longest trajectory part. This segmentation allows the system to leverage pre-acquired trajectory portions while efficiently computing only the necessary connection segments, reducing overall computational time while maintaining complete path coverage.
Solution Approach 2:
Instead of calculating the entire trajectory from scratch, the system acquires a trajectory corresponding to a similar environment and uses its longest trajectory part that lies within the moving object's accessible area. This partial action approach computes only the essential connection segments rather than the full trajectory, significantly reducing calculation time while ensuring reliability through proper connection and smoothing.
Data Source
Figure 1
Figure 2
Figure 3A~3C
AI summary
A trajectory generation device includes a storage unit that stores a plurality of trajectories; a trajectory acquisition unit that acquires a trajectory, corresponding to an environment similar to a current environment, from the plurality of trajectories stored in the storage unit; and a trajectory generation unit that calculates a longest trajectory part, which is present in a moving object moving area in the trajectory acquired by the trajectory acquisition unit, and generates a trajectory by connecting both ends of the calculated longest trajectory part to a predetermined start point and a predetermined end point respectively.