基于随机神经网络的数据关联组合优化研究

本文研究密集多回波环境下的机动多目标数据关联问题，通过对联合概率数据关联方法性能特征的分析，将其归结为一类的束组合优化问题，进而应用随机神经网络Ｂｏｌｔｚｍａｎｎ机的组合优化求解策略，结合改进的模拟增益退火方法，提出了一种新颖有效的机动多目标快速随机神经数据关联组合优化算法，克服了传统，ＪＰＤＡ存在出现的计算组合爆炸现象，仿真结果表明，该方法不仅收敛速度快，而且计算量小，关联效果好，回波愈密集，其

Investigation on Combinatorial Optimization of Data Association Based on Stochastic Neural Network

In this paper,the properties of the joint probabilistic data association(JPDA) are analyzed,andthe data association of multi-maneuvering targets is reduced to be a sort of constraint combinatorial optimization problem.Based on Boltzmann machine and simulated gain annealing ,a new algorithm called fast stochastic neural joint probabiliastic data association(FSNJPDA)is presented . The simulations show that the computation combinatorial explosion of the JPDA has been solved,and the FSNJPDA is effective and reliable .