Exercises Modernization

An Air Force Unit Develops a Training Assistance System, Giving Flight Debriefs a 'Digital-Intelligence Partner'

空军某部研发训练辅助系统，让飞行复盘有了“数智搭档”

PLA Daily (解放军报) 11 July 2026

Summary

An unidentified PLA Air Force unit stationed at a desert airfield has developed and fielded a flight training assistance system—dubbed the 'digital-intelligence partner'—that automates post-sortie debrief report generation, flags safety hazards, and converts raw flight parameter data into structured tactical assessments, with the system now in trial use across multiple aviation units. The article documents a persistent bottleneck in PLA Air Force tactical training: manual flight data interpretation was consuming pilot time that should go to tactics study, a problem acute enough during aircraft-type conversion training that it measurably extended qualification timelines. The unit-level, in-house development model described here—R&D team embedded in the training cycle, iterating algorithms overnight against same-day flight data—fits a broader PLA pattern of pushing digital and AI-enabled tools down to the unit level rather than waiting for top-down procurement, raising the question of how consistently such locally developed systems are evaluated, standardized, or integrated across the force.

Translation

An Air Force Unit Uses Technological Means to Improve Training Quality and Effectiveness—

Flight Debriefs Now Have a 'Digital-Intelligence Partner'

■ Yao Chunming, Tang Jun

Recently, deep in the desert, combat aircraft returning from a mission slowly taxied into the hangar on the runway of an Air Force unit's airfield. After training concluded, pilots made their way to the tactical discussion room to debrief the flight.

Inside the discussion room, a large screen clearly displayed the situational map of the just-completed aerial engagement, with key-node debrief points simultaneously popping up. "With the 'digital-intelligence partner' (数智搭档), debrief efficiency has improved noticeably." A pilot clicked the mouse, and on the screen, every tactical node and every instance of communication coordination from the training sortie was laid out at a glance.

The "digital-intelligence partner" the pilot referred to is a flight training assistance system independently developed by the unit.

"In the past, flight data was interpreted mainly by hand. After finishing training, pilots had to sort through the data as quickly as possible, and often worked late into the night doing so," the unit's leadership told this reporter. Sorting through data was eating into the time pilots had to study tactics, and in order to free pilots from the tedious work of data processing, they decided to develop a digital-intelligence system.

This task fell on the shoulders of a research and development team within the unit. After receiving the assignment, members of the R&D team immersed themselves in studying air combat training rules, widely solicited pilots to fill out surveys, and systematically worked through each problem the digital-intelligence system needed to solve. To identify the right technical approach, they made multiple visits to research institutes for investigation, repeatedly discussed the issues with experts in the field of digital-intelligence technology, and ultimately settled on a preliminary plan.

"This system was refined together with the pilots." Team member Ji Xinyu recalled that during development, they embedded themselves deeply in the training process and debriefed alongside the pilots. "During that period, we followed training during the day to collect requirements, optimized algorithms through the night, and then tested the improvements in training the next day—going around and around like that without a break," Ji Xinyu said.

Once the system framework was built, team members continuously refined the software—from debrief logic to interface layout, from how data was presented to how it was operated. Their efforts were not in vain: after a year of hard work, this flight training assistance system built on digital-intelligence technology was put into use.

This reporter observed during training that before a flight training sortie, a pilot need only set the core parameters in advance and the system automatically establishes an evaluation baseline. The moment a flight ends, a debrief report is generated immediately, complete with information on key nodes, points of attention, and more. Beyond that, the system can also help pilots promptly identify safety hazards and reduce flight risks.

When pilot Yu Zehao first transferred to the unit, he faced the challenge of converting to a new aircraft type. "In the past, conversion training meant spending a lot of time on data interpretation. Now with this system, I can put more energy into familiarizing myself with the aircraft and understanding tactics." With the assistance of the "digital-intelligence partner," Yu Zehao successfully completed his conversion training in considerably less time than expected.

"In the past, interpreting flight parameters relied mainly on pilot experience. Now the system turns vague subjective impressions into clear data information, making flight training more scientific and efficient," the unit's leadership explained. This digital-intelligence system has already been put into trial use across multiple aviation units and has received recognition from many frontline pilots and mission commanders. In the next phase, they will continue to expand the system's scope of application, working to ensure that digital-intelligence technology better serves the development of combat capabilities (战斗力建设).

Original Chinese

空军某部运用科技手段提升训练质效—— 飞行复盘有了“数智搭档” ■姚春明唐浚近日，大漠深处，空军某部机场跑道上，返航归来的战机缓缓滑入机库。训练结束后，飞行员走向战术研讨室复盘飞行过程。研讨室内，大屏幕上清晰呈现刚刚空中对抗的态势图，各关键节点的讲评要点也同步弹出。“有了‘数智搭档’，复盘效率明显提升。”一名飞行员轻点鼠标，屏幕上，此次训练的每个战术节点、每次通信协同一目了然。这名飞行员口中的“数智搭档”，是该部自主研发的一套飞行训练辅助系统。 “以前飞行数据主要靠人工判读，飞行员结束训练后得尽快梳理数据，常常为此忙到深夜。”该部领导告诉笔者，梳理数据挤占了飞行员研究战术的时间，为了把飞行员从繁琐的数据处理工作中解放出来，他们决定研发一套数智系统。这项任务落到了该部某研发团队肩上。受领任务后，研发团队成员深入学习空战训练规则，广泛邀请飞行员填写调查问卷，把数智系统需要解决的问题逐个梳理清楚。为了找准技术路径，他们多次赴科研院所调研，与数智技术领域专家反复研讨，最终敲定了初步方案。 “这套系统是我们和飞行员一起打磨出来的。”团队成员季鑫宇回忆，研发期间，他们深度嵌入训练过程，跟着飞行员一起复盘。“那段时间，我们白天跟训收集需求，晚上连夜优化算法，第二天在训练中检验改进成果，就这样周而复始连轴转。”季鑫宇说。系统框架搭建完毕后，团队成员从讲评逻辑到界面布局，从数据呈现方式到操作方法，不断对软件进行完善。功夫不负有心人，经过一年攻关，这套依托数智技术打造的飞行训练辅助系统投入使用。笔者跟训看到，飞行训练前，飞行员只需提前设置核心参数，这套系统便能自动建立评判基准；飞行一结束，讲评报告立刻生成，关键节点和注意事项等信息一应俱全。不仅如此，这套系统还能帮助飞行员及时排查安全隐患、降低飞行风险。调入该部之初，飞行员于泽昊面临改装新机型的挑战。“以前改装，要花很多时间在数据判读上。如今有了这套系统，我可以把更多精力放在熟悉战机和理解战术上。”在“数智搭档”辅助下，于泽昊顺利完成改装训练，所用时间比预期短了不少。 “以往判读飞行参数主要依赖飞行员经验，如今系统把模糊的主观印象变成清晰的数据信息，使飞行训练更加科学高效。”该部领导介绍，这套数智系统已在多支航空兵部队投入试用，得到许多一线飞行员和任务指挥员的认可。下一步，他们将继续拓宽系统应用范围，努力让数智技术更好地为战斗力建设服务。