基于卡尔曼滤波的动态权值融合

发布时间：2018-09-18 08:06

【摘要】：在雷达航迹融合过程中,采用多传感器测量值融合的方法能够摒除单一信息源不全面的缺点.加权平均融合为广泛使用的融合方法,但传统的权值固定的加权平均融合虽然能综合多路传感器信息,却无法自适应的根据测量值优劣倚重更有利的测量信息.因此,本文提出将固定权值改进为动态权值的融合方法,实时改变各路测量信息参与融合的权重.每次融合前,先将多路传感器测量值求简单算术平均后进行卡尔曼滤波,把滤波后的值与各路测量值作差,这相当于对传感器信息的优劣作出预判,每路测量信息的融合权值则与该差绝对值成反比.最后,通过仿真实验证明,该改进方法较之前的加权平均融合明显提高了目标的融合精度.
[Abstract]:In the course of radar track fusion, the shortcomings of single information source can be eliminated by using the method of multi-sensor measurement fusion. Weighted average fusion is a widely used fusion method, but the traditional weighted average fusion with fixed weights can integrate multi-sensor information, but it can not adaptively rely on more favorable measurement information according to the measurement value. Therefore, this paper proposes a fusion method in which the fixed weight is improved to the dynamic weight, and the weight of the information involved in the fusion is changed in real time. Before each fusion, the simple arithmetic average of the measurement value of the multi-channel sensor is calculated first, and then the Kalman filter is carried out, and the value of the filter is different from the measured value of each channel, which is equivalent to making a pre-judgement of the information of the sensor. The fusion weight of each measurement information is inversely proportional to the absolute value of the difference. Finally, the simulation results show that the improved method improves the target fusion accuracy obviously compared with the previous weighted average fusion method.
【作者单位】：四川大学计算机学院;阿坝师范学院生化系;
【基金】：国家空管科研课题(GKG201403001)
【分类号】：TN957.51

【相似文献】