Robot process
AssigneeKUKA Deutschland GmbH
InventorsManuel Kaspar, Jonas Schwinn
Life of this patent
Estimating the term…
Abstract
A method for executing a process, in particular using at least one robot, includes executing a run-through of the process, detecting a value of a first process variable, and detecting an assessment of this executed process run-through. Assessment learning steps are then repeated multiple times, wherein run-throughs of the process using varied process controls are executed and additional assessments are detected. A first quality factor model of the process, which model determines a quality factor for the process on the basis on the first process variable, is machine-learned based on the detected assessments and values of the first process variable. The method further includes repeating process control optimization steps multiple times.
Claims17
A method for executing a process on a robot, the method comprising:
- executing a first run-through of the process with a process control;
- detecting a value of a first process variable for the executed first run-through;
- and
- detecting an assessment of the executed process run-through;
- repeating assessment learning steps multiple times, the assessment learning steps comprising:
- varying the process control using an optimizer to form a varied process control on the basis of values of a quality criterion obtained from previously executed process run-throughs,
- executing a second run-through of the process with the varied process control,
- detecting a value of the first process variable for the executed second run-through, and
- detecting an assessment of the executed second run-through;
- wherein a first quality factor model of the process, which model determines a quality factor of the process on the basis of the first process variable, is machine-learned on the basis of the detected assessments and values of the first process variable;
- repeating process control optimization steps multiple times, the process control optimization steps comprising:
- varying the process control using the optimizer to form a changed process control on the basis of values of the quality criterion for executed process run-throughs,
- executing a third run-through of the process with the changed process control, and
- detecting a value of the first process variable for the executed third run-through;
- wherein the value of the quality criterion for at least one of the third run-throughs executed with one of the changed process controls is determined on the basis of a quality factor determined by the machine-learned first quality factor model on the basis of the value of the first process variable, which value is detected for the third run-through;
- and
- in response to a termination criterion being fulfilled, controlling the robot to perform the optimized process having the changed process control.
A system for executing a process on a robot, the system comprising:
- means for executing a first run-through of the process with a process control;
- means for detecting a value of a first process variable for the executed first run-through;
- and
- means for detecting an assessment of the executed first run-through;
- means for repeating assessment learning steps multiple times, the assessment learning steps comprising:
- varying the process control using an optimizer to form a varied process control on the basis of values of a quality criterion obtained from previously executed process run-throughs,
- executing a second run-through of the process with the varied process control,
- detecting a value of the first process variable for the executed second run-through, and
- detecting an assessment of the executed second run-through;
- wherein a first quality factor model of the process, which model determines a quality factor of the process on the basis of the first process variable, is machine-learned on the basis of the detected assessments and values of the first process variable;
- means for repeating process control optimization steps multiple times, the process control optimization steps comprising:
- varying the process control using the optimizer to form a changed process control on the basis of values of the quality criterion for executed process run-throughs,
- executing a third run-through of the process with the changed process control, and
- detecting a value of the first process variable for the executed third run-through;
- wherein the value of the quality criterion for at least one of the third run-throughs executed with one of the changed process controls is determined on the basis of a quality factor determined by the machine-learned first quality factor model on the basis of the value of the first process variable, which value is detected for the third run-through;
- and
- means for controlling the robot to perform the optimized process having the changed process control.
Dependencies are inferred from the claim text.