Task-Specific Robot Grasping for Diverse Objects Without Reprogramming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robotic systems face complexity in programming when handling diverse objects with varying tasks, as they require additional programming for each object and task, which can be burdensome and inefficient, especially when objects are randomly delivered and tasks differ.
Innovation Solution
A robotic system that includes a gripper movable within a 3-D volume, an articulatable portion, and an object detection system connected to a computer with a neural network capable of analyzing images to generate and evaluate grasp locations based on the next task, eliminating unsuitable positions based on the robot's movement path and available space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional robotic systems are used to handle diverse objects with varying tasks, then the robot can perform repetitive tasks, but additional programming becomes overly-burdensome and complex for each object and task
Solution Approach 1:
The robotic system performs self-programming by automatically generating task-specific parameters and grasp plans through neural network processing. The system captures images of objects, processes them through trained neural networks to determine optimal grasp locations and task parameters, and executes tasks without human programming intervention. This self-service capability eliminates the need for complex manual programming while maintaining high adaptability to diverse objects and tasks.
Solution Approach 2:
The system dynamically changes operational parameters based on object characteristics and task requirements. Neural networks process object images to generate variable parameters including grasp location coordinates, grasp force, task type, and task-specific parameters. These parameters are automatically adjusted for each object-task combination, enabling versatile operation without fixed programming.
2Reliability
If the robot is programmed to handle each object and task individually, then proper operation is assured, but the programming becomes overly-burdensome especially when objects are randomly delivered
Solution Approach 1:
The system performs preliminary actions by pre-training neural networks on large datasets of objects and tasks before actual operation. The neural networks are trained offline to recognize object features, determine grasp locations, and predict task parameters. During actual operation, the pre-trained networks quickly process new objects without requiring time-consuming programming, ensuring both reliability and efficiency.
Solution Approach 2:
The system uses image capture and processing to create digital representations of physical objects. The neural networks process these image copies to determine grasp locations and task parameters, eliminating the need for physical programming for each object. This copying approach allows the system to handle randomly delivered objects efficiently while maintaining proper operation.
3Productivity
If the robot selects grasp positions without considering upcoming tasks, then grasping can be performed quickly, but the system lacks adaptability to task requirements
Solution Approach 1:
The system dynamically determines grasp locations and task parameters based on the specific task requirements. The neural networks process object images in conjunction with task information to generate task-specific grasp plans. This dynamic approach allows the system to adapt grasp selection to upcoming tasks while maintaining high productivity through automated processing.
Solution Approach 2:
The system uses feedback from task requirements to adjust grasp location selection. The neural networks receive task information as input and generate grasp plans that are optimized for the specific upcoming task. This feedback mechanism ensures that grasp selection is both efficient and task-appropriate, balancing productivity and adaptability.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A robot operable within a 3-D volume includes a gripper movable between an open position and a closed position to grasp any one of a plurality of objects, an articulatable portion coupled to the gripper and operable to move the gripper to a desired position within the 3-D volume, and an object detection system operable to capture information indicative of the shape of a first object of the plurality of objects positioned to be grasped by the gripper. A computer is coupled to the object detection system. The computer is operable to identify a plurality of possible grasp locations on the first object and to generate a numerical parameter indicative of the desirability of each grasp location, wherein the numerical parameter is at least partially defined by the next task to be performed by the robot.