Robotic Depalletization Using Height-Based Layer Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic depalletization systems face inaccuracies and failures when dealing with stacks of objects of different sizes and shapes, particularly in partial depalletization scenarios, as they struggle to accurately identify and separate objects within mixed SKU stacks.
Innovation Solution
A controller-implemented method for robotic depalletization that detects pallet layers and objects using imaging data, determines the pallet height difference, and updates layer or SKU index data to facilitate partial depalletization based on user-defined thresholds, enabling precise identification and removal of objects in layer-based or SKU-based modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional robotic depalletization systems are used on stacks of objects of different sizes and shapes, then the system can perform basic depalletization operations, but system inaccuracies and failures occur when attempting partial depalletization of mixed SKU stacks
Solution Approach 1:
The system segments the stack into distinct pallet layers by detecting height differences between objects. The controller identifies layer boundaries by comparing the height of the first object (defining the top layer) with subsequent objects, dividing the mixed SKU stack into manageable layers that can be selectively depalletized. This segmentation enables accurate partial depalletization by targeting specific layers rather than treating the entire stack as a uniform entity.
Solution Approach 2:
The system performs partial depalletization by removing only the top pallet layer or specific portions of the stack based on user-defined thresholds. Instead of depalletizing the entire stack, the controller compares pallet height differences to thresholds to determine when a layer has been fully removed, enabling precise control over the depalletization process and allowing the system to stop after removing only the necessary portion of the stack.
2Ease of operation
If the system attempts to identify and separate objects in mixed SKU stacks, then partial depalletization becomes possible, but system inaccuracies and failures increase due to object variability
Solution Approach 1:
The system transitions from two-dimensional object recognition to three-dimensional height-based layer detection. By measuring pallet height differences in the vertical dimension and comparing object heights to identify layer boundaries, the system creates a more robust method for distinguishing objects in mixed SKU stacks. This dimensional approach reduces inaccuracies by relying on height differences rather than attempting to identify and separate individual objects with varying sizes and shapes.
3Measurement precision
If layer-based depalletization is implemented with height threshold comparison, then precise layer identification is achieved, but the system complexity increases
Solution Approach 1:
The controller continuously monitors pallet height during the depalletization process and compares measured height differences to predefined thresholds. This feedback mechanism enables the system to dynamically determine when a pallet layer has been fully removed and adjust operations accordingly. The threshold-based feedback loop provides precise layer detection while maintaining relatively simple control logic, as the system only needs to compare height measurements against stored threshold values rather than performing complex object recognition algorithms.
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
Various embodiments are directed to methods, apparatuses, systems, computing devices, computing entities, and the like for robotically depalletizing objects. In various embodiments, a controller-implemented method for robotic depalletization may comprise detecting a first object of a plurality of palletized objects arranged in a plurality of pallet layers, the first object defining at least a portion of a top pallet layer; identifying an initial pallet height defined by the first object; identifying a second pallet height defined by a second object of the plurality of palletized objects; detecting a pallet height difference based at least in part on the second pallet height; and determining that the top pallet layer has been depalletized based at least in part on a comparison of the pallet height difference to a threshold pallet layer height.