基于模糊神经系统的多传感器数据融合算法  

Neural-Fuzzy-Based Multisensor Data Fusion Architecture

作　　者：荣健[1] 乔文钊[1] 

机构地区：[1]电子科技大学物理电子学院,成都610054

出　　处：《电子科技大学学报》2010年第3期376-378,424,共4页Journal of University of Electronic Science and Technology of China

基　　金：国家自然科学基金(60572079)

摘　　要：将自适应模糊神经推理系统(ANFIS)和卡尔曼滤波器应用于目标跟踪系统中,构成多传感器数据融合算法。该算法假设在目标运动过程中,过程噪声和测量噪声是相互独立的高斯白噪声序列。使用ANFIS分别对目标的加速度和测量噪声的方差进行估计,通过卡尔曼滤波器获得目标后验状态,最终由神经网络对多传感数据进行融合得到系统输出。仿真结果表明,该算法可以通过自适应调整跟踪参数有效地防止目标丢失。An adaptive neural-fuzzy-based multisensor data fusion architecture for target tracking systems is presented.In this architecture,both process noise and measurement noise are modeled as uncorrelated zero-mean Gaussian noise sequences.Adaptive-network-based fuzzy inference systems（ANFIS） are employed to detect and estimate target maneuvers and measurement noise covariance matrices.They are considered as an adaptive mechanism to cooperate with Kalman filters to process multiple sensor data,which are fused by a specific neural network to obtain optimal results.The results of simulation demonstrate this architecture can avoid mistracking effectively by adjusting tracking parameters.

关 键 词：数据融合 模糊逻辑 卡尔曼滤波 多传感器系统 神经网络 

分 类 号：TN193[电子电信—物理电子学]