华中科技大学学报（自然科学版）2017，Vol.45Issue(9)：11-17,7.DOI:10.13245/j.hust.170903
GPU加速的高超声速目标双通道IMM跟踪算法
Dual-channel IMM tracking algorithm for hypersonic target with GPU acceleration
摘要
Abstract
An improved dual channel IMM (interactive multi model)algorithm was proposed.The likelihood function of the prediction error was represented independently using the edge distribution of x and y-dimensional residuals.The filter model was matched with the target motion in different di-mensions.Firstly,the mapping relationship of residual-likelihood function-model probability updating was analyzed,and the feasibility and validity of the algorithm were proved.Then the parallel compu-ting platform of GPU (graphics processing unit)was introduced,and the channel level-model level embedded parallel scheme was designed to improve the real-time performance.Simulation results show that the proposed algorithm is more accurate than the traditional IMM algorithm based on CPU platform,the prediction error of the proposed algorithm is better than that of the real noise likelihood distribution,and the real-time error and tracking error are smaller.关键词
目标跟踪/交互多模型(IMM)/高超声速飞行器/图形处理器(GPU)/双通道/并行运算Key words
target tracking/interacting multiple model (IMM)/hypersonic vehicles/GPU (graphics processing unit)/dual channel/parallel operation分类
信息技术与安全科学引用本文复制引用
李凡,熊家军,毕红葵,兰旭辉..GPU加速的高超声速目标双通道IMM跟踪算法[J].华中科技大学学报（自然科学版）,2017,45(9):11-17,7.基金项目
国家自然科学基金资助项目 (61401504). （61401504）
