Hemodialysis machines have an important role in the support and maintenance of life functions, wherein the strict control of peristaltic pumps is a serious need. The current paper targets to present soft computing methods, which can be applied for the peristaltic pumpcontrol problem, demonstrating their applicability on real systems as well. The performances of these controllers are compared with those of the classical PID controller, in order to highlight the advantages and applicability of these modern control strategies over classical controllers. An adaptive fuzzy-logic system and an adaptive neuro-fuzzy inference system were designed to fulfill the proposed control criteria. The requirements for the sought controller were formulated in three aspects: to be fast, to eliminate the residual error, and to be insensitive to disturbances. The designed controllers were compared and the one demonstrating the best performance has been tested on a real system as well.
ASJC Scopus subject areas
- Theoretical Computer Science
- Computational Theory and Mathematics
- Artificial Intelligence