www.intralogisticsmag.com
11
'25
Written on Modified on
End-to-End Robotic Picking Brings Fully Unmanned Warehousing Closer
Geekplus introduces a robot arm picking workstation combining generalizable AI perception and AMR coordination to automate high-volume operations with diverse SKU inventories.
www.geekplus.com
Reference image
A new automated picking workstation for high-volume logistics environments
Warehouse automation has traditionally been limited by the difficulty of automating the picking stage, especially when handling large and diverse SKU inventories. Geekplus has introduced a new Robot Arm Picking Station, designed to work alongside its autonomous mobile robots (AMRs) to complete the last major manual process in goods-to-person workflows. With this launch, the company moves from partially automated picking lines to end-to-end unmanned picking, creating a foundation for fully autonomous warehouse facilities.
Core innovation: a generalizable robotic picking model
At the center of the new workstation is Geek+ Brain, a general-purpose perception and manipulation model trained on extensive real-world warehouse data. The system is engineered to recognize and grasp tens of thousands of SKUs—including deformable, reflective, and irregularly shaped goods—without the need to retrain models for each product category. This addresses one of the most persistent constraints in robotic picking: adapting to heterogeneous item inventories in high-throughput settings.
Performance engineered for industrial logistics
The workstation delivers three key technical advantages:
- Fast deployment and modularity
The system’s modular architecture allows installation within 48 hours, with support for scalable capacity increases to match seasonal peaks or packaging changes. It is designed to be “plug-and-play” within existing AMR-driven workflows.
- High throughput and continuous operation
Model-level optimization enables millisecond-scale inference, supporting coordinated arm and AMR workflows and achieving throughput several times higher than manual picking. Features like double-pick verification and integrated scanning reduce error rates and maintain consistent cycle times under 24/7 workloads.
- On-premises data processing
All inference and data management take place locally, avoiding cloud transmission and aligning with stringent enterprise data-security requirements.
A new era of unmanned warehouses.
Integration across logistics platforms
The Robot Arm Picking Station integrates seamlessly with Geekplus’ AMR ecosystem, enabling:
- Single-order picking for rapid order responsiveness
- Batch picking for throughput optimization in high-volume operations
This flexibility supports sectors with varied order-profile structures, including e-commerce, retail distribution, pharmaceutical fulfillment, and 3PL warehouse networks.
Toward fully unmanned warehouses
The new workstation extends Geekplus’ unified software and robotics architecture, reducing integration complexity between picking, movement, and storage solutions. With robotic packing technologies under development, Geekplus is progressing toward fully unmanned warehouse environments, where all stages—from storage retrieval to picking and packing—operate autonomously within coordinated multi-robot systems.
www.geekplus.com
Toward fully unmanned warehouses
The new workstation extends Geekplus’ unified software and robotics architecture, reducing integration complexity between picking, movement, and storage solutions. With robotic packing technologies under development, Geekplus is progressing toward fully unmanned warehouse environments, where all stages—from storage retrieval to picking and packing—operate autonomously within coordinated multi-robot systems.
www.geekplus.com
