The Chinese Academy of Sciences and Shield AI cooperated to complete the autonomous flight verification of Jinfeng III UAV

Chinese Academy of Sciences Jinfeng III UAV. The Chinese Academy of Sciences and the American company Shield AI jointly completed the autonomous flight verification of three "Jinfeng III" UAVs a month ago. (File photo)

[Reporter Tu Jumin/Taipei Report] my country continues to expand the development of military unmanned vehicles. Shield AI, an American company that specializes in artificial intelligence and unmanned vehicles, teamed up with the Chinese Academy of Sciences to successfully complete the autonomous flight verification of three "Jinbee III" drones a month ago. Chen Yiru, Shield AI's Asia Government Relations Director, said that the cooperation between the two parties "hit it off immediately", and after entering the engineering stage, the integration work was completed in only about three months. In practical applications, AI pilots lead detailed flight operations, allowing operators to concentrate on mission judgment and decision-making.

The Chinese Academy of Sciences and Shield AI signed a memorandum of cooperation in February this year. The two parties announced on August 6 this year that three Chinese Academy of Sciences "Jinfeng III" vertical take-off and landing drones, using Shield AI's Hivemind technology, have completed a practical flight demonstration of collaborative autonomous reconnaissance, take-off and landing missions. The key to this cooperation is the Hivemind artificial intelligence autonomous flight software developed by Shield AI, which can be regarded as the "AI pilot" of unmanned vehicles.

Zeng Guoguang, president and co-founder of Shield AI, said in a written statement that the key to improving defense resilience is not simply to increase the number of hardware equipment. Taiwan already has quite mature capabilities in hardware manufacturing. However, without independent system support, many existing equipment will be unable to function once GPS or communications are interfered with or interrupted. Hivemind has begun to introduce Taiwan's local platform and has made good progress. Local engineers from the Chinese Academy of Sciences have successfully completed the construction of AI pilots on a platform made in Taiwan, allowing Taiwan to truly master its independent capabilities.

Please read on...

Exclusive interview with Max Chen, Director of Government Relations, Asia Region, Shield AI. (Photography by reporter Zhang Jiaming)

Chen Yiru said in an interview that the Hivemind autonomous system can allow drones and unmanned boats to perform tasks in environments with severe electrical warfare interference, GPS and communication restrictions. Originally, multiple drones may need to be controlled by multiple operators. Now, the operator only needs to give instructions at the base to "control multiple drones at once" and perform tasks such as large-scale reconnaissance and search.

Regarding the actual application of the "Jinbee III" UAV, Chen Yiru gave an example. Among the three "Jinbee III" drones, one may serve as a communication host node, and the other two may perform reconnaissance missions, which can expand the search area. In addition to reducing the operator's cognitive load, in high-pressure battlefield environments, AI can also be used to lead detailed flight operations, allowing the operator to focus on mission judgment and decision-making.

Talking about this cooperation with the Chinese Academy of Sciences, Chen Yiru described it as "hit it off immediately." He said that the software engineers of the Chinese Academy of Sciences have excellent capabilities, and the two teams can cooperate to handle the hardware architecture of different drones, connect command and control (C2) systems, process intelligence surveillance and reconnaissance (ISR) signals, and deploy the software to edge computing hardware. From software simulation testing, hardware loop testing, to actual flight testing and public performances, it takes about 3 months to complete.

Another Hivemind application result of Shield AI is its cooperation with Leihu, a private unmanned vehicle manufacturer. Chen Yiru said that engineers from both parties started with the circuit and hardware circuits and introduced Hivemind into the Thunder Tiger unmanned boat. At the end of July, the unmanned boat autonomous group tour and other maritime demonstrations were conducted in Dapeng Bay, Pingtung.

Chen Yiru said that Taiwan's unmanned vehicle industry is quite mature, and he hopes that Shield AI's Hivemind system will help the domestic drone industry play a "functional acceleration" role, while also cultivating Taiwanese software engineers and improving independent research and development capabilities.