神经网络在多目标分布式融合中的应用研究

Application of Neural Network in Multi-sensor Multi-target Distributed Fusion

提出了适合分布式多传感器多目标 (MSMT)数据融合估计算法 ,对这种推广的分层融合估计算法进行了数学推导和描述 ,研究了利用神经网络进行MSMT跟踪融合的网络结构和算法 ,进行了计算机模拟仿真。仿真结果表明神经网络融合估计可提高跟踪精度 ,并且这种融合方法是有效的。

In the paper, a new method of data fusion in multi sensor multi target(MSMT) tracking, i.e. a generalized algorithm for hierarchical fusion, which is especially suited to the distributed MSMT tracking system,is presented and described. A neural network structure and algorithm for the MSMT tracking fusion is studied. The simulation of the neural network method is discussed. Simulation results show that the neural network method is valid and can improve the fusion accuracy.