Supervisory control theory (SCT) provides a formal approach to supervisory controller synthesis. In this paper, SCT is used to design a supervisory controller for a patient support system. This system is used to position a patient in a Magnetic Resonance Imaging (MRI) scanner. To improve the evolvability of the design, the uncontrolled system and the control requirements are modeled independently, using small, loosely coupled and minimally restrictive automata. An implementation of the synthesized supervisor is realized by means of a transformation to an automaton in the Compositional Interchange Format (CIF). The supervisor is validated by means of hardware-in-the-loop simulation, using the real patient support table.