Decanting and Mixed Load Depalletizing for Warehousing Applications

Discover how machine learning (ML) and next-generation artificial intelligence (AI) vision technology can unlock new applications in the warehouse of the future.

Labor challenges continue to impact the growing warehousing and e-commerce markets; AMT engineers have identified best-in-class technologies that can use automation to solve complicated material handling challenges in these industries.  Warehouse automation involves the complexity of combining material handling equipment, sensor technology, conveyance, and software to move materials at very high speeds. The warehouse of the future includes more efficient picking and sorting technology with fewer errors and increased reliability, regardless of the packaging. Today’s systems integrate with leading-edge AI software and next-generation machine vision for autonomous, accurate product handling.

The Labor Challenge – Upskilling the Workforce

Adding automation solutions in the warehousing and e-commerce markets can positively impact labor productivity.  With training, employees will transition from mundane tasks (dull, dangerous and dirty) to value-added work, helping them enjoy a more fulfilled life.  Improved job satisfaction reduces employee turnover. Operator safety is improved by removing stress injury applications, such as heavy lifting and repetitive tasks.  Reducing the manual labor in warehousing and e-commerce applications leads to increased throughput and reduced per-unit handling costs.

Depalletization: How Hard is it to Pick Up a Box?

Depalletizing, the action of unloading a single item, box, case, or layer from a pallet, takes place at the front of line. This is sometimes known as “induction,” where the unloaded material is fed into another storage system or process. While the action seems simple, there are many challenges that a robot system must overcome to accurately and effectively pick up materials off a pallet.

There are three types of pallet categories in a robotic depalletizing solution: homogeneous, heterogenous, and rainbow pallets. In addition to pallet types, there are several different kinds of depalletizing to consider: single case and multi case depalletizing, row depalletizing, and full layer depalletizing.  The holy grail of depalletizing, mixed-load depalletizing (unloading boxes of different sizes, shapes and weights), has inherent challenges due to the variety of product to be unloaded.

Automated Depalletizing in the Warehouse

Warehouse automation is a growing trend, largely impacted by the difficulty of finding and retaining employees. To handle materials at high speeds, warehouse automation requires the intricate integration of materials handling equipment, sensor technology, conveyance systems, and specialized software to overcome operational challenges.  The warehouse of the future includes more efficient picking and sorting technology with fewer errors and increased reliability, regardless of the packaging.

Automating the process of case and pallet induction offers multiple benefits for the warehousing industry, including a reduction in manual labor requirements and costs, a reduction in the cost of handling per unit, an increase in operator safety by removing repetitive stress injuries, and growth in the scale of operation.

But, automating this process is no walk in the park; it takes experts who know what to expect from this type of project. These challenges highlight the importance of selecting the right integrator to take on the project. AMT has extensive experience successfully integrating depalletizing systems in warehousing settings for our clients.

Typically in depalletizing applications, there are five top challenges to address:

1. Case size variability
2. Printed matter/graphics
3. Pick pattern optimization
4. Pick tool or dunnage or slip sheets
5. New case introduction – Volume of SKU’s - probabilistic or machine learning

Along with these challenges, there are 10 key considerations for developing the optimal depalletizing solution.

Decanting in Warehouse Applications

Decanting is a term most often used in the wine industry, referring to the act of gradually pouring a liquid from one container to another (typically without disturbing the sediment). In warehousing depalletizing applications, decanting is the act of using machine vision technology and next-generation artificial intelligence (AI) vision to plan a sequence of moves to empty a container.  Decanting software instructs a robot to pick the optimal number of boxes in a single pick that will fit into a defined tote size or cubic space. Specifically for distribution or fulfillment center applications, the decanting process is defined as moving product from single SKU pallet loads to single SKU storage totes. A homogeneous decanting process performs a multi-pick with optimized pick count to maximize tote fill and minimize the number of moves to complete a layer of a single, uniform type of product.

AI and Machine Learning in Decanting Applications

To be able to tackle the problems faced with replacing repetitive labor in decanting applications, AI and ML technology is a must in order to optimize the number of boxes per pick and minimize the total number of picks.
Behind the scenes in decanting software, machine learning essentially creates a lengthy yet simple equation that accepts inputs and calculates outputs. The equation has a lot of constants, and programmers randomize/adjust the constants until the output is satisfactory

Once sufficiently trained, you can give the system a new input that it has never seen before and the equation is tuned to accommodate and give the correct output. At this point, it has "learned" an equation that solves the initial problem and now it can be used to find patterns and relationships in the inputs that are difficult to define with conventional programming.

Paired with a properly set up vision system, the application can now tackle problems that would be expected of a person on the line. It can “see” what it is looking at on the pallet and with AI technology, the system makes decisions of what to pick and how.

Decanting System Details and Considerations

In a robotic decanting system, an overhead 3D vision system finds all the boxes and determines where to place to tool for a best pick.  Sensor technology and user-friendly software determine exactly where the boxes, bags, or other product containers are, as well as their size and orientation.  Armed with this information, the robot can move product quickly and with great accuracy.

Extra constraints to be considered:

• The group of boxes to pick cannot be larger than the destination tote
• The end of arm tooling (EOAT) must be placed where it won't pick up additional unwanted boxes
• The EOAT always goes to the center of the tote, so the boxes shouldn't overhang outside the EOAT

Programming for the above considerations will create a functional system, but further optimizations can help with throughput, such as having the system plan a few picks ahead to minimize the number of moves required.

Start Your Adoption of Depalletizing Automation

For more information on palletizing and depalletizing, read our reference guide.

If you’re interested in learning more about the benefits your facility could experience from automating your depalletizing processes, be sure to reach out.