Robot Object Handling Through Learned Action Databases
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object handling procedures for robots in automated production and assembly lines are complex, require specific programming for each task, lack flexibility, and incur high costs due to continuous reprogramming needs.
Innovation Solution
A process involving a robot with a graphical user interface, sensors, and a learning phase to recognize and interact with objects, using a computer to create an action database for flexible and efficient handling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If specific programming languages are used to instruct the robot for each single action, then the robot can perform tasks with high precision, but the management of production processes becomes complex and error-prone
Solution Approach 1:
The system creates a digital model (action database) that copies and stores information about object characteristics and robot actions. Instead of programming each action separately, the system uses stored action models that can be retrieved and executed, simplifying management while maintaining precision through the detailed digital representations.
Solution Approach 2:
The system performs preliminary actions by pre-recording and storing action information in the action database during a learning phase. This allows the robot to later execute tasks by retrieving pre-prepared action sequences rather than requiring complex real-time programming decisions.
2Loss of time
If graphical user interfaces are used to program robots, then programming speed is improved, but the procedures remain complex and require expensive robots
Solution Approach 1:
The robot performs self-learning by autonomously observing human demonstrations and automatically creating action models in the database. This eliminates the need for complex graphical programming interfaces and expensive specialized robots, as the system teaches itself through observation and replication of human actions.
Solution Approach 2:
The system replaces complex graphical user interface interactions with a learning-based approach where the robot observes and replicates actions. This substitution of traditional programming mechanisms with machine learning reduces complexity and cost while maintaining ease of use.
3Productivity
If traditional handling procedures are used, then robots can perform repetitive tasks, but they lack flexibility and require continuous reprogramming for changes
Solution Approach 1:
The system transitions from static pre-programmed sequences to dynamic adaptive behavior. The robot uses the action database to learn and adapt to new objects and movements by observing demonstrations, allowing it to maintain high productivity while gaining flexibility to handle changes without continuous reprogramming.
Solution Approach 2:
The system changes the fundamental parameter of how robot instructions are defined - from fixed programming to learned action models. By storing and retrieving action information in the database, the robot can adapt to different objects and movements by selecting and modifying stored action sequences rather than requiring complete reprogramming.
Data Source
AI summary
A process of interacting with objects is provided including a robot with an end effector, a driver of the end effector and sensors for acquiring at least one environmental or interaction parameter; a computer for controlling the robot; an instruction block configured to instruct the robot to move the end effector according to instructions defined by the instruction block.