基于模糊卡尔曼滤波的多传感器焊缝偏差检测

Detection Method for Welding Seam Deviation Based on Fuzzy Adapted Kalman Filtering

在先验知识未知的情况下,为准确检测焊缝水平偏差信息,提出一种基于模糊卡尔曼滤波的多传感器焊缝偏差检测算法.采用基于模糊逻辑的卡尔曼滤波算法实时处理焊缝传感器观测的数据,构建2个模糊推理控制器分别在线调整系统噪声和测量噪声的协方差系数,以使得系统噪声和测量噪声逼近真实噪声,再将滤波后的值进行自适应动态加权融合得到准确的焊缝偏差值.最后,将该方法应用于焊缝偏差检测系统,仿真结果验证了算法的有效性.

In order to detect the welding seam horizontal deviation accurately in case of unknown prior knowledge, a welding seam deviation algorithm based on the fuzzy Kalman filter and multi-sensor information fusion was proposed. The algorithm of Kalman filter based on fuzzy logic was adopted to process observation data of sensors in real-time. The fuzzy inference algorithm was used to adjust the weight coefficients of process noise covariance matrix and mea- surement noise covariance matrix to make model noise gradually close to the true noise level. Then using dynamic weighted data fusion algorithm fuse the fuzzy Kalman filter value, hence, the final weld deviation was obtained. Fi- nally, simulation results showed that MSDF in weld seam deviation detection is genuine and effective after it is ap- plied to the detection system of welding seam deviation.