基于IMM的大尺寸回转零件动态位置检测方法

Dynamic position detection method for large rotary parts based on IMM

为了解决大尺寸回转零件动态位置不易准确测量或测量成本较高等问题,提出了一种大尺寸回转零件动态位置实时检测的新方法。该方法根据大尺寸回转零件实际运动规律,构建了匀速运动模型和"当前"统计模型,进而给出了基于两种运动模型的交互多模(interacting multiple model,IMM)算法。该方法将多个光栅传感器检测的动态位置信息分别通过IMM算法计算得到融合值,然后运用滤波误差协方差阵对各融合值进行最优加权估计。仿真表明,基于IMM算法和最优加权估计算法的动态位置检测方法具有较小的位置误差。将该方法应用于某型火箭整流罩卧式铆接设备空心轴转角定位系统进行工程验证,实测结果表明该方法具有较高的位置检测精度和稳健性。

It was difficult to measure the dynamic position for large rotary parts accurately and the cost of measurement was high. However, to avoid these problems, according to the law of actual motion of large rotary parts, this paper presented a no- vel real-time detection method, which built constant velocity model and ＂current＂ statistical model, and then established in- teracting multiple model （1MM） algorithm based on the two models. Every dynamic position variable detected by multi-grating sensors was computed by the IMM algorithm to get the final fusion value, and the final fusion values were weighted estimated by using filtering error covarianee matrix. The method based on IMM algorithm and optimal weighted estimation algorithm had a small position error, which was demonstrated by Monte Carlo simulation. And the method was also used to measure the angu- lar position of the hollow shaft of a rocket fairing horizontal riveting apparatus. The experimental result shows that the method has a high position detection accuracy and robustness.