基于CNN的位置融合模型

A CNN-Based Model of Position Fusion

卷积神经网络(CNN)在图像处理领域取得了巨大成功,受其启发,提出一种基于CNN的目标位置融合模型,并介绍模型的网络结构、数据预处理方法,以及数据集生成方法。针对多站基于目标测向信息的位置融合场景进行仿真,通过对原理和试验结果的分析,表明其在具备异构传感器的复杂场景中也将具备较强的适应能力。

CNN achieves great success in image processing.Enlightened by that,a CNN-based model for position fusion is proposed,and its network structure,data preprocess method,as well as the method of the dataset generation are introduced.The simulation of multi-station positioning fusion scenario based on DOA information is carried out.It is shown by the principle and the simulation results that the model could be adaptable well even in complex scenario with heterogeneous sensors.