With the advancement of technology and the development of artificial intelligence, the civil aviation industry is undergoing a profound intelligent transformation. The launch of airport intelligent robots, smart travel managers and other products in visible places to consumers has made their travel more convenient; In invisible places, a complete domestic civil aviation operation system is being planned and developed, and computing power companies are actively researching and launching domestic chips to contribute to ecological development. Intelligent applications ensure passenger travel. At the China Eastern Airlines booth at the 3rd CATA Aviation Conference, the new generation intelligent service robot "Xiaodong" attracted many visitors to stop and watch. The reporter learned that this intelligent robot applied to airport terminal scenery has been integrated into the large model Q&A knowledge base and check-in system. When passengers check in, they only need to touch and click on the robot to handle one-stop procedures such as boarding, refunds, and changes. Southern Airlines has launched the Southern Airlines Smart Travel Manager to address pain points such as tight schedules and frequent changes for business travelers. Based on a high-quality dataset covering billions of customers, the platform utilizes multimodal big model technology in a scenario based manner to create a one-stop service platform that can recognize passenger intentions and link travel needs through natural language dialogue. The travel manager also has special features such as one click import of video guides and intelligent obstacle avoidance for airport navigation. The relevant person in charge of China Southern Airlines told reporters that the travel planning scenario library of this product currently covers more than 2000 scenic spots across the country. In addition to intelligent interactive products aimed at passengers, civil aviation enterprises have also launched intelligent products that ensure the safety and efficiency of civil aviation operations throughout the entire chain, including equipment inspection and maintenance, luggage transportation, cabin sound monitoring, and other operational processes, with the assistance of AI for component maintenance and safety supervision. According to He Feng, General Manager of the Data Department of China Southern Airlines, in order to promote the intelligent construction of scenarios, China Southern Airlines has carefully laid out 196 innovative scenarios from three aspects: flight operation, passenger experience, and management support. 74 of them have been successfully put into production, including flight risk radar, China Southern Airlines navigation notification assistant, and apron operation safety supervision. Taking the safety supervision of apron operations as an example, apron safety management has always been a traditional problem. The old supervision mode has narrow coverage, high labor costs, and lagging supervision. With the development of AI technology, China Southern Airlines' data technology has divided 16 scenarios on domestic aprons involved in the company, deployed nearly 18000 cameras and more than 4000 vehicle monitoring hosts, collected real-time operation images, applied AI intelligent analysis to automatically judge dangerous operation behaviors, reminded on-site operators in real time, synchronously triggered safety supervision intervention mechanisms, and increased the coverage of apron dangerous operation supervision to 85%. The average recognition time of violations has been shortened to less than 1 second, and the response time to accidents has been improved by more than 90%. Cheng Kai, a trainer from the Training Management Department of Hainan Airlines Aviation Technology Co., Ltd., introduced to reporters the company's independently built VR mechanical skills training application system and demonstrated how to wear VR glasses for the B787 aircraft wrap around inspection project. He stated that there are numerous training programs and complex scenarios involved in the field of mechanical maintenance. Projects such as disassembly and assembly of hydraulic mechanical components, replacement of slide packages, etc. still pose certain safety risks. Through the VR mechanical skills training application system, students can form a cognitive understanding of relevant scenarios in advance, thereby improving the quality and efficiency of training. Cheng Kai stated that the current system has covered 5 aircraft models including A320, B787, E190, etc., and has completed nearly 20000 training sessions. Sun Qiongwei, General Manager of AVIC Digital Products Division, told reporters that a large amount of manpower is needed in the operation of civil aviation. The means of informatization and artificial intelligence can not only add another layer of firewall to data system security, but also greatly release manpower, reduce the burden on operators, and improve the level of security. Domestic computing power joins hands to build the "smart brain" of civil aviation. Han Jun, Deputy Director of the Civil Aviation Administration of China, stated at the 3rd CATA Aviation Conference that during the "15th Five Year Plan" period, the development of civil aviation will be more prominent in terms of quality and efficiency improvement, functional system optimization, and resource allocation efficiency improvement. The entire industry will coordinate high-quality development and high-level safety, continuously enhance the systemic and industrial resilience of civil aviation, and improve the independent innovation and industrial integration development system of civil aviation. In 2024, China Aviation Information Technology Co., Ltd. and Air China officially reached a strategic cooperation, forming a strong alliance in the core operation field of airlines to create a new domestically produced operation control system for China's civil aviation. This also marks China Aviation Information Technology Co., Ltd.'s entry from the passenger service field to the aircraft operation field. ”Sun Qiongwei stated, "Currently, both sides are concentrating their advantageous resources to jointly tackle the two core systems of dispatch and flight planning. In the next step, they will continue to focus on key core technology areas such as navigation, intelligence, meteorology, and aircraft performance to ensure the safety of critical operational data and ensure that the safe operation of China's civil aviation is in their own hands." To support the massive civil aviation operation system, domestic computing chip companies and civil aviation actively cooperate to help build a smart civil aviation ecosystem. Yang Hongxuan, Vice President of Haiguang Information, told reporters that "the demand for advanced computing power in civil aviation is huge. Traditional general computing power and intelligent computing power are needed in multiple scenarios such as aerodynamic simulation, digital twin, flight prediction, etc., to achieve safe, green and intelligent development of civil aviation." Several industry insiders mentioned that in the development process of civil aviation industry, chip level cooperative development based on existing application scenarios is an important way to build a good civil aviation ecosystem. Guo Lin, Deputy General Manager of Beijing Jietong Huasheng Technology Co., Ltd., stated that the cooperation between civil aviation enterprises and computing power enterprises should not only be limited to finding suppliers when projects are implemented, but should also start with early support cooperation from the process of chip product research and development and solution layout, so as to be more relaxed when the application scenarios mature. In response, Haiguang Information has launched a new product form of AI super cluster solution, based on AI computing open architecture design, which can help ecological partners and users obtain high-level technical sharing, deep customization optimization, and professional technical support in AI application storage performance acceleration, liquid cooled super node system design, and GPU framework development. Yang Hongxuan stated that the system is designed based on an open AI architecture, fully compatible with multiple brands of AI acceleration cards, and fully compatible with mainstream AI software ecosystems. It can create a powerful computing foundation for more scenarios such as training and inference of civil aviation models, fine-tuning of industry models, and development of multimodal models. The computing power distribution in the civil aviation industry is scattered, and the computing power architecture is also very diverse, often with multiple heterogeneous computing powers being used in parallel. In the future, Haiguang Information hopes to assist civil aviation users in achieving diversified computing power integration and intelligent scheduling, and help build a network for civil aviation computing power scheduling. ”Yang Hongxuan said. (New Society)