关键词: 飞机-跑道系统, 动力学响应, 长短期记忆网络（LSTM）, 智能代理模型

Abstract: To solve the problem that the dynamic simulation of the aircraft-runway system mainly relies on numerical mod鄄 els, resulting in low computational efficiency and making it difficult to meet the real-time response requirements of the intelligent runway system platform, this paper proposes an intelligent surrogate model for the dynamic re鄄 sponse of the aircraft-runway system based on long short-term memory (LSTM). Taking the B737-800 aircraft and Jinan Yaoqiang International Airport runway as the research objects, the model achieves fast and accurate prediction of dynamic response. Based on the multi-condition dataset constructed by extending the coupled dy鄄 namics solver, a dual-layer LSTM architecture is designed and the training strategy is optimized to enhance the model′s generalization ability. The results show that, compared with the traditional numerical simulation, the av鄄 erage relative error of the model prediction under the steady taxiing condition is less than 3%; at a taxiing speed of 260 km/h, the computation time for a single run is reduced from 277.20 s to 0.07 s, and the computational effi鄄 ciency is improved by approximately 3 960 times, with good generalization performance. This model provides a new technical path for the real-time and accurate analysis of the dynamic response of the aircraft-runway system.

Key words: aircraft-runway system, dynamic response, long short-term memory (LSTM), intelligent surrogate model