en.Wedoany.com Reported - X Square Robot, an embodied intelligence company based in Shenzhen, livestreamed a logistics parcel induction automation demonstration on X and YouTube on August 12. With the coordination of its self-developed WALL-B AI foundation model and a high-performance 6-axis robot arm, the system identified, grasped, and organized parcels piled haphazardly after truck unloading, feeding them one by one onto a conveyor belt for scanning and automated sorting.

Parcel induction is a labor-intensive "3D" (dirty, dangerous, and dull) task in logistics sorting, with high manpower demands and high turnover rates, making shift scheduling for night shifts and seasonal peaks particularly difficult. As labor costs rise, automated equipment capable of continuous operation offers a potential path to improved productivity and return on investment.

The technical challenge in this stage lies in the fact that the shape of the parcel pile changes with every grasp: boxes and soft packages press against each other, labels face the wrong direction, and the optimal grasping surface is often buried. At unloading docks, this disorder makes parcel induction a bottleneck for throughput and one of the most labor-intensive steps in sorting.

Wang Qian, founder and CEO of X Square Robot, stated that the problem logistics automation needs to solve is not whether a robot can complete a single clean grasp, but whether it can consistently make correct decisions as the parcel pile changes, recover when problems arise, and keep the rest of the operation running continuously. The company designs automation around real-world work, not idealized environments.

WALL-B is the company's self-developed embodied AI foundation model, responsible for evaluating each parcel and its surrounding environment, determining how to handle it, and adjusting strategies as conditions change. Rather than relying on manually programmed rules for each parcel type, it enables the robot to handle unexpected situations in the physical world—during the livestream, the 6-axis robot arm dynamically intervened to rescue a parcel that had gone off course.

At the execution level, the high-performance 6-axis robot arm handles the work. Its multi-angle range of motion allows the system to approach parcels from different directions for grasping, flipping, flattening, and repositioning. Before parcels reach the conveyor belt, the robot arm can turn labels toward the scanner, smooth out soft packaging, and square up boxes, thereby improving barcode scanning and reducing returns and manual handling caused by poorly positioned parcels.

Rather than requiring a single robot to handle all tasks, X Square Robot adopts a specialized approach, matching hardware and AI to the actual needs of the induction process instead of forcing the workflow to accommodate the robot. The result is a precise combination of speed, accuracy, and cost. For X Square Robot, this demonstration means more than just moving parcels—it reflects a broader vision for embodied intelligence: making AI practical in the real world, enabling robots to perceive, reason, and act in complex physical environments.

X Square Robot is a company developing general-purpose embodied intelligent robots with high-precision manipulation capabilities. It is building foundation models for the physical world, integrating models, robot hardware, and a scalable, model-driven high-quality data pipeline system. Its robots perform long-horizon tasks involving perception, reasoning, and action in scenarios such as home and daily environments, scientific research and education, logistics and warehousing, and industrial operations.