The special‑operation track of this contest contains two competition items, intelligent manipulation of high‑voltage power‑distribution switchgears and intelligent fire‑scene reconnaissance. Jiuzhou‑Software dispatched the Jiuzhou Light‑Chasing Team and the Jiuzhou Electric‑Guard Team for the two assignments separately.

"Opening the cabinet door is merely the initial step, and the genuine test lies ahead." At the testing site, Wang Xingxing, the company’s contest participant, kept his eyes fixed on the surveillance screen and explained operational procedures to journalists. The Xuan‑Jia robot travelled autonomously toward a 2.2‑meter‑tall power‑distribution switchgear, extended its robotic arm steadily, gripped the cabinet‑door handle accurately and pulled the door outward. Every movement was completed neatly, and the red‑and‑green indicator lights inside the cabinet lit up afterwards.

After rapid identification, the robot pressed the corresponding button precisely. It then picked up the roughly 2‑kilogram cylindrical key from the designated spot and inserted it firmly into the 4‑centimeter‑wide keyhole. The whole sequence of motions was executed seamlessly.

According to Wang Xingxing, Xuan‑Jia is positioned as an integrated special‑purpose device for electric‑power inspection and manipulation, with its core competency being full‑process autonomous fine‑grained operation. Per competition specifications, the robot needs to complete the entire switch‑operation workflow including cabinet‑door opening, indicator‑light recognition, button operation and precise key insertion‑and‑removal within 25 minutes. "This 10‑kilovolt high‑voltage cabinet carries substantial hazards for manual operation, and human workers cannot stay on standby for equipment maintenance around‑the‑clock," Wang Xingxing remarked. "Essentially, the robot takes over the most dangerous human‑performed tasks."

Tense testing also went on at the adjacent emergency‑reconnaissance zone. Equipped with infrared thermal imaging, 3D LiDAR and a multi‑sensor fusion perception system, the Sentinel robot carried out fire‑site reconnaissance missions inside a simulated smoke‑filled, low‑light setting. The competition task required it to penetrate smoke, traverse bumpy road sections and steep slopes, locate survivors and transmit on‑site data in real‑time within 15 minutes. "Visibility drops close to zero amid heavy smoke, which imposes extreme loads on its vision, locomotion and communication systems," a training‑site technician explained.

When asked why these two tough challenge categories had been selected, Wu Hongjun, product lead for embodied‑intelligence hardware at Jiuzhou‑Software, shared three major considerations. "Electric‑power switching‑operation and fire‑site reconnaissance represent high‑risk, essential‑demand scenarios. Human rescuers face considerable casualty risks, and robot replacement constitutes an inevitable trend for the unmanned transformation of the industry."

He further stated that Jiuzhou‑Software possesses years‑long expertise in AI‑driven embodied intelligence, multi‑sensor fusion and intelligent‑control algorithms. The firm is accelerating the construction of a full‑chain unmanned safety solution covering inspection, manipulation, reconnaissance and early warning, targeting the two core sectors of unmanned electric‑power maintenance and urban emergency rescue.

"We avoid conceptual‑only products and exclusively develop deployable, combat‑ready special‑purpose equipment," Wu Hongjun commented. After the contest, relevant technological achievements will be rapidly applied to new‑type power‑system construction and urban emergency‑safety frameworks. This rigorous competition will also serve as an important window for observing Mianyang’s innovation strengths within embodied intelligence and special‑purpose equipment industries.