" "

Automated pallet pickup? Here’s what most solutions get wrong

Automated Pallet Detection

About MOV.AI 

MOV.AI disrupts Autonomous Mobile Robot development with a Robotics Engine Platform that contains everything needed to quickly build, deploy and operate intelligent robots.

The AMR Revolution: Materials on the Move 

It’s no secret that Autonomous Mobile Robots (AMRs) are reshaping intralogistics and warehouse operations. According to Interact Analytics, 120,000 AMRs were shipped in 2022, a number that is expected to grow by 45% in 2023 and by more than 30% in 2024.

Of these, nearly half of the robots were for material transport – largely pallet movers and cart tuggers.

The industry is embracing automation to deal with labor shortages, the need for speed, and more. However, on the path towards realizing potential efficiencies, there are critical points that can pose a challenge – especially in existing facilities with dynamic mixed environments where humans and robots collaborate. One of these is pallet pickup and drop-off. 

The cost of pallet pickup errors

Automated pallet pickup and drop-off are complex tasks that require numerous different capabilities. Issues with pallet pickup can have profound implications on warehouse efficiency, causing delays and escalating costs:

  • Route Interruptions: When an AMR misidentifies a pallet, it may deviate from its planned route, causing delays in delivering goods or completing tasks.
  • Improper alignment may cause an AMR to get stuck, or worse, cause damage or safety hazards.
  • Delays and interruptions can have a cascading effect, impacting subsequent tasks and schedules. These directly translate to lost productivity.

Automated Pallet Pickup Challenges

Let’s consider for a moment an automated pallet pickup process: The AMR needs to autonomously navigate and travel to the pallet’s location, identify the correct pallet, position itself properly to pick it up, place its forks securely beneath it, lift the pallet safely and securely, and maneuver the pallet (which can weigh a ton) out of its position. All this, in a minimal amount of time and while avoiding any damage to the pallet, its cargo, the surrounding environment, and itself.

Autonomous Pallet Movers’ ability to perform this task is impacted by:

  • Pallet Variations: Pallets come in various sizes (EU, US, Half a pallet), different materials (wood, plastic, metal), and even condition (damaged, wrapped in plastic, dirty, etc.). These variations challenge AMRs to adapt their detection algorithms for each type.
  • Lighting Variations: Uneven or changing lighting conditions in warehouses can confuse AMR sensors, making it difficult to accurately identify pallet edges and dimensions.
  • Technology Limitations: Current sensor technologies like LiDAR and cameras have limitations in terms of range, resolution, and field of view, which can restrict accurate pallet detection, especially in cluttered environments. Algorithm accuracy and the ability to calculate fine maneuvers also play a role in perception and motion planning.
  • The physical environment: The amount of space needed and that is available for maneuvering and precision movement.
  • The human factor: Humans, by nature, introduce disorder. They may place the pallet a few centimeters off the mark, making it impossible for some pallet detection technologies to identify it and pick it up.

What’s wrong with the common techniques for automated pallet pickup?

To ensure smooth pallet pickup, AMR manufacturers and automation integrators have embraced a range of technologies designed to ensure accurate pallet placement in a precise location that is optimized for pallet detection, pickup, and maneuvering.

Floor Markings and Guides:

Physical markings or guides on the warehouse floor can be used to indicate proper pallet placement. This helps human operators or autonomous systems to align pallets correctly. For example, colored lines or patterns on the floor can guide pallet placement.

Pros: Defines a pickup spot that is optimized for pickup success.

Cons: Requires human operators to be highly accurate, which demands extra effort and time from employees. This impacts their efficiency, and placement errors still occur. Flexibility to make changes is also reduced, as any change requires adjustment to markings.

Physical Solutions:

To avoid human error, some warehouses use automated or semi-automated pallet positioning systems that ensure pallets are placed in predefined locations. This may involve conveyor systems, laser guidance, robotic palletizers, dedicated racks or other equipment that assists in aligning and organizing pallets.

Pros: Accurately placed pallets and a low error rate.

Cons: High investment costs and a rigid setup that limits the ability to make changes, impacting operational and business flexibility.

So what’s the problem?

The problem with these solutions is that instead of solving the problem, they provide external workarounds.
Instead of improving AMRs’ ability to detect and pick up pallets in a “natural” warehouse environments operated by humans, they “fix” the environment and the behavior of the people in it. These external fixes come with a price tag of capital investment and loss of operational flexibility.
Advanced AMRs that can perform high-precision pallet detection, alignment, and placement with no external aids are a game changer for continuous, efficient, and flexible warehouse operation.

 

What if your AMR could detect pallets like a human?

High-precision detection, alignment, and maneuverability
MOV.AI has introduced a new, highly accurate pallet perception and alignment algorithm into its Robot Engine Platform™. Combined with high maneuverability and advanced routing and fleet management capabilities, MOV.AI allows AMRs to pick up pallets from almost anywhere, with no need for environmental adjustments, route markings, or even placement markings.

Built into our platform, these algorithms equip AMR manufacturers with the tools to create highly autonomous pallet robots that seamlessly integrate with existing business processes and minimize disruption.

At the heart of our solution is the combination of two algorithms: advanced pallet perception and alignment algorithms. This powerful combination delivers:

  • Higher Picking Success Rate: The detection algorithms go beyond simple identification, pinpointing the precise location and orientation of each pallet. This enhanced perception, coupled with precise alignment maneuvers, dramatically reduces mis-picks and ensures seamless, efficient operation.
  • Robustness to Imperfections: Our system understands that the real world isn’t always perfectly controlled. Inherent inaccuracies in pallet placement or AMR positioning are accounted for, ensuring reliable performance even in less-than-ideal scenarios.
  • Operational and Business Flexibility: AMRs can pick pallets from almost anywhere, directly from the floor, and with no need for external infrastructure. This maximizes operational flexibility and adaptability.

 

Key Steps in the Process:

Goto Scan Keypoint:

    • Navigation guides the AMR to designated scan key points.
    • From these positions, the AMR scans the area for available pallets to pick up.
    • Decision-making logic can be based on inputs from Warehouse Management Systems (WMS), operator commands, or pre-defined rules.

Pallet Detection:

    • Pallet detection runs in the expected area using a 3D camera signal.
    • Pose estimation estimates the location and orientation of the pallet.
    • Pallet classification identifies the type of pallet (EU/US/Half-US), useful when the pallet type is not known in advance.

Pallet Alignment:

    • Using the estimated pose, a path is calculated.
    • Navigation controls the AMR to align with the pallet.
    • Once aligned, the AMR positions itself under the pallet for pickup.

Flexible operation in any layout

MOV.AI pallet perception and alignment can detect pallets placed freely in an area. It can also identify pallets in a grid-based layout, a common mode of operation in manual warehouses. Support for grid-based layouts allows warehouses and employees to continue operating familiarly with no need to adapt processes to the new autonomous equipment. This intuitive approach eliminates the need for drastic workflow changes, ensuring a smooth transition to automation.

Pallets in grid layout
Pallets in grid layout

Here is an example of an AMR running MOV.AI pallet detection solution aligning to a pallet that was manually displaced from its picking area.

 

Simple implementation using the MOV.AI Robotics Engine Platform

The pallet detection solution is fully open and available for simple integration into AMR software via MOV.AI Flow™. 

Robot manufacturers can use the perception algorithms as-is or adapt them to their needs: Customize logic and communication to add pallet types, create APIs, or change inputs and outputs as needed for specific AMRs and use cases. 

Editing pallet pickup in MOV.AI Flow™

 

Final thoughts

As AMR adoption grows, warehouse operators, integrators and manufacturers are gaining a better understanding of what it takes to enable seamless operation, maximize AMR utilization and unlock expected efficiency gains.

Pallet detection is a seemingly minor activity that can make it harder to achieve automation and efficiency goals. Most common solutions fall short as they require environmental adjustments, necessitate changes to existing processes, and reduce flexibility.

MOV.AI’s pallet scanning, detection, and alignment capabilities can offer a high precision solution that provides AMRs with true autonomy through:
Seamless Automation Integration: Minimal disruption to existing workflows for a smooth transition.

  • Grid-Based Efficiency: Familiar layout mirrors manual operations for ease of adoption.
  • Unmatched Precision: Advanced detection and alignment algorithms for higher picking success rates.
  • Real-World Robustness: Accommodates inaccuracies in pallet placement and AMR positioning.
  • Business and Operational Flexibility: Removing the need for purpose-built infrastructure and predefined routes reduces the need to change existing processes during initial deployment and enables changes down the line to meet business needs.

 

Contact us to hear more.

 

Sign up for news update

*By providing my details I agree to the MOV.AI Terms and conditions and Privacy policy. I also agree to receive emails from MOV.AIand I understand that I may opt out of MOV.AI subscriptions at any time.

Ready to build awesome robots?

We use cookies to so that you can enjoy the best experience on our website. If you continue to use this site we will assume that you are happy with it.
Skip to content