System identification based on orthogonal polynomial regression analysis method

By way of orthogonal polynomial regression to fitting step response curve, the type and order of the controlled object are identified through the ratios of time corresponds to three specific points on the curve; relations between 2~8-order same volume object step response curve inflection point coordinate, the time-constant and the order n are given. For same volume object, the time-constant can be obtained by making the fitted curve´s second derivative to zero. For first-order with Pure lag time link, s-shape step response curve but will be expressed by first-order with Pure lag time link and higher-order non-same volume object, proposes the method to make the linear fitting with the orthogonal polynomial regression to get the time-constant. Simulation test shows, the method is simple, fast and accurate to identify the two types objects.