Robot Pose Selection for Faster Multi-Position Task Execution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for configuring and controlling robots fail to efficiently utilize multiple possible starting poses for tasks like drilling or riveting, leading to suboptimal performance and increased cycle time.
Innovation Solution
A method and system that determine and select optimal target poses for robots based on a predefined set of possible starting poses, using user-input-dependent criteria to configure the robot arrangement and environment for improved task execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a robot uses a single predetermined starting pose for a task, then the configuration is simple, but the cycle time increases and task efficiency decreases
Solution Approach 1:
The patent pre-calculates and stores multiple possible starting poses for each task in a database before the robot actually executes the task. This preliminary preparation allows the robot to quickly select from pre-evaluated poses during task execution, reducing cycle time without adding complex real-time calculation requirements to the robot controller.
Solution Approach 2:
The system dynamically selects the most appropriate starting pose from multiple predetermined options based on task-specific criteria such as cycle time optimization, collision avoidance, and robot configuration. This dynamic selection capability allows the robot to adapt its starting position to each specific task instance, improving efficiency without requiring a completely flexible real-time pose generation system.
2Loss of time
If multiple possible starting poses are considered for each task, then cycle time is reduced and task efficiency improves, but the configuration and control complexity increases
Solution Approach 1:
Multiple starting poses and their associated quality criteria are pre-calculated and stored in a database before task execution. This preliminary action transfers the computational complexity from real-time robot control to an offline setup phase, allowing the robot controller to simply retrieve and select from pre-evaluated options during task execution, thus reducing cycle time without increasing real-time control complexity.
Solution Approach 2:
The patent introduces a database as an intermediary between the task requirements and the robot execution. This database stores pre-calculated starting poses and their quality criteria, acting as a mediator that provides the robot controller with ready-to-use information without requiring complex real-time calculations, thereby reducing cycle time while maintaining simple real-time control.
3Productivity
If optimal target poses are determined based on user-input-dependent criteria, then task execution is optimized, but the system complexity increases
Solution Approach 1:
The system optimizes task execution by changing parameters such as starting pose selection, target pose determination, and path planning based on user-input-dependent criteria. These parameter changes are applied to pre-calculated options from the database, allowing optimization of task execution efficiency without requiring complex real-time control algorithms, as the optimization is achieved through selective parameter application rather than complex computation.
Data Source
Figure 1~2
AI summary
A method according to the invention comprises the step of: ascertaining (S110) at least one possible desired pose of a robot (10) of a robot arrangement for at least one prescribed initial position (Xa) and task of the robot on the basis of a selection set ({{R}i}), ascertained particularly before operation of the robot, of possible initial poses (Pa, ij) prescribed for this at least one task and initial position (Xa, i). Additionally or alternatively, the method comprises the steps of: ascertaining (S30) a starting set of initial poses (Pa, ij) of the robot for the at least one initial position (Xa, i) of the robot; selecting (S70) at least one initial pose (Pa, ij) from the starting set on the basis of the at least one task of the robot and a prescribed selection criterion (Gij), particularly one prescribed on a user input basis; and adding (S80) this initial pose to the selection set as a possible initial pose prescribed for this initial position and task.