In Makura, a subsidiary of Zhenhuo Shenzhou Holdings (00861), Kejie and the famous robot WAIC launched the world's first intelligent cluster

The Zhitong Finance App learned that on July 17, during the World Artificial Intelligence Conference (WAIC), a special live display attracted widespread attention from the industry. Currently, most of the scenes of embodied intelligent robots active in the media are dances, martial arts performances, etc., which are still far from continuous operation, complex collaboration, and stable delivery in real production environments. And robots that work in real industrial and commercial environments and can produce production value are the focus of the industry's attention.

This time, Kejie, a smart supply chain brand under Shenzhou Holdings (00861), teamed up to extend the camera from the WAIC booth to the Tianjin Wuqing Kejie Logistics Warehouse. Multiple intelligent robots continued to carry out picking operations in the actual logistics warehouse, using actual production orders to complete the entire process from task reception, route planning, product identification and handling, to sorting truck positioning and complex packaging, to intuitively show the progress of physical intelligence from “booth presentation” to “real work”.

Inside the warehouse, multiple wheeled intelligent robots shuttled through Kejie's actual operating warehouse of thousands of square meters to independently complete the entire warehousing and selection process. Multiple robots operate in parallel within the same channel, and are automatically avoided and queued through system scheduling and path algorithms; when the power level falls below the preset threshold, the robot automatically goes to the recharge point, and the rest of the equipment continues to operate during charging of some devices to ensure business continuity.

It is worth noting that the robot does not completely replace labor in the warehouse, but works with Kejie frontline employees to complete order fulfillment. The robot completed the selection process, and Kejie staff reviewed it. Through human-robot collaboration, it not only leverages the advantages of stable robot operation, but also preserves the key role of humans in anomaly identification, quality judgment, and order delivery.

Unlike proof of concept in a laboratory environment, the core of this presentation is “real warehouse” and “reality.” As a 5A logistics company with nearly 200 warehouses and a daily peak of 5 million orders, Kejie made it clear from the beginning that robots must not only do demonstration picking; they must enter real warehouses and continue to work with real orders.

To achieve this goal, Kejie opened a thousand-square-meter real commercial warehouse in Wuqing, Tianjin to the Big Name Robotics Project Team. This is not a temporary simulated site, but a real operating warehouse carrying the daily deliveries of many leading brands. Robots must face positioning drift in a long corridor environment, handling and adaptation at different shelf heights, extreme differences in products from a single item to a full box, and diverse packaging in different categories such as clothing and accessories — these are all key questions that are difficult to replicate in the laboratory, but determine whether the robot can be used commercially.

More importantly, Kejie has transformed more than 20 years of supply chain operation experience into robot operating rules. Which channels and locations should the robot take, where to set up avoidance points, in which areas should the sorting truck be placed, and under what conditions can an empty truck be called, and when to recharge to ensure uninterrupted operation — these SOPs and on-site flow line designs for warehouse operations were determined by the Kejie team based on an understanding of actual business processes, and embedded in the robot's operation logic. The optimal path for the robot to walk combined with the order position is automatically determined by the system, giving priority to completing tasks that are closer and reducing ineffective round trips. When operating in a multi-machine cluster, multiple robots do not “lock up” each other when they meet, but instead dynamically adjust routes according to the task location and channel status, giving priority to orders with a closer path.

In the field optimization of the main body and action plan, Kejie and the big name robot co-created to deeply collaborate with the main body and scene process to form a realistic and intelligent warehousing selection solution.

The robot's initial robotic arm design continued the solution previously tested in small scenarios such as pharmacies, but problems were discovered in the actual operation of Kejie's logistics warehouses: when there were very few items in some bins, there was a risk that the original degree of freedom could not be absorbed or the posture was limited. Based on field operation feedback, the robot customization selected for this scenario increased the degree of freedom of the robotic arm, changing from the 7 degrees of freedom initially used to 8 degrees of freedom, increasing the ability to move the wrist.

In terms of handling strategies, the robot originally tried to work with left and right hands at the same time to improve efficiency, but it actually interfered with the SKU distribution and operation flow lines of different customers in the warehouse. In the end, the robot selection scheme was adjusted to: for small SKUs such as wrist braces and headscarves common in sports brands, the system will match the single suction nozzle scheme to prevent the double suction nozzle from absorbing two items at the same time; for clothing categories with larger bags, double suction nozzles are used to ensure stability. The end of the arm is equipped with a suction nozzle device, and the other side is equipped with a dexterous hand to stably tow the sorter.

The opening of orders and systems also reflects Kejie's system capabilities. The Shenzhou Treasury supply chain software platform independently developed by Kejie is deeply connected to the robot scheduling system — after the consumer places an order, the order is automatically transferred to the Shenzhou Treasury through the OMS system. The system automatically allocates tasks suitable for the robot to perform operations. The robot independently receives and executes them, and there is no need for manual intervention on an order-by-order basis throughout the process.

This joint launch of WAIC is a concentrated expression of Kejie's “scenario-defining technology” concept. It shows the deep meaning of Kejie's position as a “technology-driven industry supply chain expert”: technology is not an isolated device or algorithm, but must be embedded in a collaborative system of orders, warehouses, and people to accept continuous verification of actual business. Kejie provides not only a real warehouse, but also incorporates more than 20 years of supply chain operation experience, standardized warehousing processes, real order data, and systematic operation rules into robot training, and participated in continuous optimization from warehouse maps, operation flow lines, order distribution to multi-machine scheduling, handling movements, and human-robot collaboration, helping autonomous intelligent robots move from “being able to operate” to “being able to operate stably in real logistics scenarios”.