Generalisation of conditioning technique for anti-windup compensation

For linear controllers, the anti-windup compensator based on conditioning technique is often treated as a special case of the generalised anti-windup compensator. The conditioning technique is then defined simply by a special observer in the generalised anti-windup compensator. Instead, in the paper the concept of back calculation or the modified setpoint is emphasised, which is indeed the conceptual basis for the derivation of the conditioning technique. The behaviour of the modified setpoint is investigated by use of a simple example, and an inherent weakness in the conditioning technique is demonstrated. A simple modification to the conditioning technique that reduces the effect of short-sightedness is presented, and its anti-windup compensation properties are investigated. Also, a general approach to the conditioning technique is proposed by introducing a filtered setpoint. This approach provides a parameterisation to deal not only with the short-sightedness problem but also with the other drawbacks of the conditioning technique. An important outcome of this generalisation is that it unifies and interprets through the conditioning technique a class of AWCs described in the literature.