Automatic testing of control systems - past, present, and future

The paper reviews the methods whereby automatic testing may be used for assessing the health, quality assurance, and fault diagnosis characteristics of control systems. The level of sophistication of automatic testing is shown to progress in parallel with developments in digital computers. Five generations of automatic test equipment are identified. The starting point is seen to be fixed-sequence tape-controlled devices, and now the fifth ATE generation anticipates closedloop mission simulation and the use of artificial-intelligencebased diagnostics. Both linear and non-linear control system testing is described in detail. The problem of fault diagnosis is complicated in analogue systems by poor signal-to-noise ratios, and significant component tolerances for nondefective parts. Consequently there is a trade-off curve for diagnosability which depends heavily on system and signal uncertainties. `Intelligent¿ signal processing and diagnostic routines greatly enhance ATE capability in this respect.