Reviewer: NEIL MUNRO article which introduces many new conceptual ideas and how these might be implemented in future control systems. It is extremely well illustrated, and uses a specific example to show how an adequate CACE Tool should not only support the user in the design of the proposed intelligent control architecture but should also be able to evaluate by simulation the resulting intelligent control system and generate the low-level computer code required for real-time realization. The third article (46 pages), by M. Jamshidi, R. Marchbanks, K. Bisset, R. Kelsy, S. Bough and D. Barak, is on the Computer-aided Design of Fuzzy Control Systems: Software and Hardware Implementations.
The material here is presented in five parts, dealing with (i) a brief introduction to fuzzy logic and fuzzy control; (ii) a brief description of two fuzzy software programs, one implemented in the symbolic computational environment Mathematics illustrating fuzzy logic principles and the other being a simulation of an inverted pendulum using Togai's Fuzzy-C expert system and the MATLAB simulation environment; (iii) two fuzzy logic case studies, one being the fuzzy control of a thermal system by NeuraLogic Corporation and the other being the fuzzy logic control of a road-traffic intersection (where comparisons with conventional and expert system controllers are given); (iv) some experimental results obtained from the real-time fuzzy control of a robotic conveyor belt and a separate laser-tracking system with particular emphasis on the effect of different computer architectures; and (v) future research directions including a combined fuzzy and expert system control. The fourth article (30 pages), by Grantham K. H. Pang, addresses Knowledge-based Control System Design. A very useful survey of expert systems for CACSD is given with copious references and a novel knowledge environment is described which supports both symbolic and numerical processing. Expert system control of multivariable systems is illustrated using a very nice systematic yet high-level design approach. There is a strong emphasis throughout this contribution on interactive design and the designer's intuition. The resulting package, called MEDAL, integrates a generic knowledge base development environment with a MATLAB-derived control environment. Various knowledge representation frameworks are considered.