舰船电子工程2025，Vol.45Issue(12)：46-50,89,6.DOI:10.3969/j.issn.1672-9730.2025.12.010
基于储备池计算的动态轨迹实时预测算法
Real-time Dynamic Trajectory Prediction Algorithm Based on Reservoir Computing
摘要
Abstract
Real-time prediction of the trajectories of moving targets using neural network models has extremely high applica-tion value.Taking the dynamic trajectory prediction of drone as a case study,this paper studies the trajectory prediction of moving targets based on the reservoir computing algorithm.The reservoir network is trained with the acquired historical of drone,and the network model generated by the training is used to predict the future movement trajectory of drone.Based on the network model trained with 5 000 frames of historical movement trajectories of drones,real-time prediction of the next moment trajectory is achieved.According to the evaluation criteria for trajectory prediction,the experimental results show that when predicting the trajec-tories for the next 10 frames,the prediction errors of 80％of the prediction points can be controlled within 0.8％.关键词
储备池计算/动态轨迹预测/无人机Key words
reservoir computing/dynamic trajectory prediction/drone分类
信息技术与安全科学引用本文复制引用
杨志良,毛煜茹,姚金杰,安建平..基于储备池计算的动态轨迹实时预测算法[J].舰船电子工程,2025,45(12):46-50,89,6.基金项目
国家自然科学青年基金项目(编号:62401518)资助. （编号:62401518）
