In new energy vehicle factories with a high enough degree of automation, what else can robots do?
Text | Fu Chong
Editor | Su Jianxun
The year 2025 is regarded as a crucial year for the commercialization of embodied intelligence. Some enterprises have used live broadcasts of industrial on - site operations to demonstrate the possibility of implementing embodied intelligence in factory transfer scenarios.
On July 9, 2025, Zhiyuan Robotics and Annu Intelligence conducted a three - hour live broadcast at the Mianyang Factory of Fuling Precision Industry. During the live broadcast, the self - developed A2 - W general embodied robot of Zhiyuan Robotics continuously carried out a two - shift task of carrying more than 800 turnover boxes per shift.
During the 3 - hour live operation broadcast, the A2 - W made 0 mistakes. Through dual - arm collaboration and high - precision operation, the A2 - W can grab, carry, and place turnover boxes of different sizes and weights. When the pallets and bins are not stacked properly, it can also independently identify and adjust its posture to grab. When an abnormality occurs during the operation, it can also correct errors independently and continue to run.
In addition, the A2 - W also demonstrates multi - modal perception and dynamic obstacle avoidance. The A2 - W can identify people, turnover boxes, and other obstacles in real - time, independently plan paths, adjust or pause operations, and has a safety redundancy mechanism when sharing the operation space with people.
Li Pengcheng, the vice - chairman of Fuling Precision Industry, said when evaluating the on - site performance of the A2 - W in the factory: "It meets the requirements of the scenario." According to Li Pengcheng, previously, the transfer work in Fuling Precision Industry mainly relied on manual labor. The work of carrying boxes was high - frequency and large - quantity, with the weight ranging from 5 kilograms to 30 kilograms, which was extremely unfriendly to human labor. The application of robots can alleviate the problem of difficult recruitment.
The "working" and carrying scenario of the A2 - W this time
However, although the A2 - W shows intelligence and accuracy in the above scenario, its efficiency still needs to be improved. Currently, the carrying cycle time (the interval time between two consecutive identical tasks) of the A2 - W is about 40 seconds, slower than the 30 - 35 - second cycle time of humans.
"Automation in factories has already been fast and good. Why do we still need embodied intelligence? Because traditional industrial robots are fixedly deployed and have poor adaptability, making it difficult to meet the flexible production needs of multiple varieties and small batches. For example, the production line iteration speed of new energy vehicles is significantly faster than that in the era of fuel - powered vehicles. The application of embodied intelligence can avoid frequent upgrades and transformations of the automated production line." Wang Chuang, the president of the general business department of Zhiyuan Robotics, explained the importance of embodied intelligence in factory production to 36Kr.
Wang Chuang also introduced that compared with automation, embodied intelligence has a stronger ability to resist interference and handle abnormal situations. When the production volume increases, its cost will also be reduced.
Fuling Precision Industry, the destination of the A2 - W's "work" this time, is a listed company in the automotive parts segment. In recent years, new businesses of Fuling Precision Industry, such as new energy vehicle electric drive systems, thermal management systems, lithium battery cathode materials for new energy vehicles, and key components of wire - controlled suspension systems, have grown rapidly, and the business of intelligent robot electric joints and components has also been added.
This operation uses the robot body and base model of Zhiyuan Robotics, and the engineering deployment is provided by Annu Intelligence. Shanghai Zhiyuan New Technology Co., Ltd. (Zhiyuan Robotics) and Fuling Precision Industry are the shareholders of Annu Intelligence.
Annu Intelligence provides solutions for the last mile from "laboratory technology to industrial implementation" of embodied intelligence. According to Yang Zeng, the engineering algorithm director of Annu Intelligence, in the subsequent scenario iterations, Annu Intelligence will build a full - dimensional engineering solution data collection system covering robot sensor data, actuator status, and overall machine operation parameters.