Multi-Arm Robot Coordination for Simultaneous Order Picking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robots in intralogistics are unable to solve complex order-picking tasks, particularly those requiring recognition of unplanned situations and simultaneous grasping and pushing of items, which are typically handled by humans.
Innovation Solution
A multi-arm robot system with a base frame, multiple manipulators, and a control unit that evaluates real-time data from acquisition units to perform complex picking tasks by recognizing situations and executing appropriate operations stored in databases, including manipulations and movements, to autonomously handle items in various conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional robots are used for order picking, then simple sequential tasks can be performed efficiently, but complex tasks requiring simultaneous grasping and pushing cannot be solved
Solution Approach 1:
The robot system is divided into multiple independent arms, each capable of performing different operations simultaneously. This segmentation allows one arm to grasp items while another pushes items aside, enabling complex order picking tasks that cannot be performed by a single robot arm sequentially.
Solution Approach 2:
Multiple robot arms are combined into a single coordinated system with shared control and perception. The arms work together as an integrated unit, with the control unit coordinating their actions to perform complex tasks such as simultaneously grasping and pushing items, thereby increasing adaptability without proportionally increasing overall system complexity.
2Productivity
If multiple robot arms are arranged in series to perform tasks in parallel, then task capacity increases, but coordination complexity and control difficulty increase
Solution Approach 1:
Each robot arm is designed with universal capabilities to perform multiple types of operations (grasping, pushing, holding). This multi-functionality allows the arms to be dynamically assigned different tasks based on real-time needs, increasing productivity while simplifying control compared to specialized arms requiring complex coordination protocols.
Solution Approach 2:
The system employs real-time feedback from sensors and vision systems to continuously monitor the state of items and arm positions. This feedback enables the control unit to dynamically adjust arm movements and coordination, managing the complexity of parallel operations through adaptive control rather than rigid pre-programming.
3Speed
If a single robot arm performs tasks sequentially at high speed, then simple picking tasks are efficient, but complex tasks requiring simultaneous operations cannot be completed
Solution Approach 1:
The single robot arm is segmented into multiple independent arms, each capable of operating simultaneously at high speed. This allows different arms to perform different operations (grasping, pushing, transferring) at the same time, maintaining high overall picking speed while enabling complex simultaneous operations that a single sequential arm cannot perform.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Storage and order picking system (12) for carrying out a complex order picking task (76) in which several items (18) must be manipulated simultaneously, comprising: a detection unit (28) configured to cyclically detect at least the items (18) with respect to their position and orientation and to generate corresponding data; and a multi-arm robot (10) comprising: a base frame (46); a plurality of arms (20) that are independently movable simultaneously and that are attached to the base frame (46); a plurality of manipulators (54), each of the arms (20) being connected to one of the manipulators (54);and an arm control unit (74) configured to move the arms (20) into arm-specific positions and orientations depending on the situation, so that several of the arms (20) simultaneously and jointly perform the complex picking task by manipulating the objects (28) with the corresponding manipulators (54) according to a predefined operation, and configured to cyclically determine the positions and orientations of the objects (18) relative to the robot (10) from the data of the acquisition unit (28) and/or from situational data.