机电一体化系统设计4(离散控制).pptVIP

Lecture 4Discrete time control Lecture outline Discrete time control introduction, sampling of continuous signals Pure discrete control design Discrete time approximation of a continuous time control design Writing code Continuous vs discrete time control Discrete time control Assumptions/consequences New discrete time design issues Two design concepts (continuous, discrete, combination of both) Selection of the sampling rate(s) - a key issue; performance vs cost. Due to the sampling, filtering of sensor (and control) signals is often required; that is, pre- and post sampling filters. Computer implementation or e.g. cascaded control loops may lead to multi rate sampling to be handled by the real-time implementation. Software implementation; what does a control algorithm look like in terms of program code (programming a dynamic system). Analysis of the effects of the transformation of the control design from continuous to discrete time. Numeric interp. of discrete/continuous Sampling continuous time signals Frequency folding Numeric example of aliasing Pre- or antialias filtering Real signals are not band limited. High frequency components must be filtered away to avoid aliasing. For digital/discrete sensors the aliasing problem is less aggravating. The prefilter is typically implemented as an analog filter with resistors, capacitors and an operational amplifier. The filter should be taken into account in control design. NEVER IMPLEMENT A FEEDBACK LOOP WITHOUT ANTIALIASING FILTER ON ALL AD-CONVERTERS Choice of sampling rates single rate systems high sampling rate is costly the frequency should be set in relation to the fastest dynamics in the closed loop characteristics (i.e. bandwidth, rise-time) of the feedback,observer or model following. or 4-10 samples per rise time Sampling interval selection based on freq. Lecture outline Discrete time control introduction, sampling of continuous signals Pure discrete control design Discrete time approximation of a con