Main directions of the research
- LMI-based feedback regulator design via TP transformation
- Fuzzy control
- Identification of peristaltic pumps
- Classical control methods
- Adaptive neuro-fuzzy systems
A detailed description of the research
In haemodialysis machines, peristaltic pumps are responsible for the fluid flows (such as blood, dialysate or ultrafiltration flow). However, the transfer volume may highly vary due to the inaccuracy of the loaded tube segment. This research focuses on the control problem related to peristaltic pumps to create accurate fluid transport.
The accurate transfer is important from more aspects: Peristaltic pumps are responsible for drug delivery. For the safety of the patient, the recommendations of medical staff (especially doctors) have to be kept, otherwise, the drug infusion is unattended. On the other hand, the fluid balance of the patient is crucial, as people on haemodialysis machines have decreased urine secretion (if any). In order to keep the total fluid volume of the patient at an optimal level the added and removed fluid volumes have to be known, which again requires an accurate transfer.
In haemodialysis machines, occurs several control problems. One of these problems is the maintenance of the fluid balance of the patient. This industry-supported research solves a complex problem that started with identification until the implementation of a developed controller.
To find the most suitable control method different control methodologies were tested (even on real machines). Controllers were designed with classical approaches (PID), soft computing methods (fuzzy control, ANFIS) and LMI-based regulation.