Snake Skeleton Graph: A New Method for Analyzing Signals That Contain Spatial Information

Abstract

Many modern engineering systems use multiple sensors for monitoring, diagnosis and control. Some of these sensors contain not only time domain and frequency domain information, but also valuable spatial domain information to which little attention has been paid. This paper presents a new method for capturing the spatial characteristics of the sensor signals. The basic idea is to model the spatial information of the signals using a Bézier surface. For example, given m one-dimensional force signals: X1(t),X2(t),[[ellipsis]],Xm(t), at each time instance t, a Bézier surface can be constructed, which describes the distribution of the force. Furthermore, lining up the surfaces at different time t1,t2,[[ellipsis]],tn, will show how the force changes as a function of time. Since the graph looks like a snake skeleton, the new method is called the snake skeleton graph. The paper first describes how the snake skeleton graph is constructed using a demonstration example: a foot walking on a plate. Then it presents an application for fault diagnosis in sheet metal stamping operation. Future research topics are also discussed.