It is usual that we are able to describe the control strategy in natural language. Our approach enables using natural language directly without thinking of the way how is the proper fuzzy control realized. Thus, the user should not think of shapes of fuzzy sets, used operations, inference engine, etc., as it is otherwise usual in fuzzy control widely considered elsewhere. We speak about linguistically oriented fuzzy control and the corresponding system is called Linguistic Fuzzy Logic Controller (LFLC). We may also see LFLC as if “partner” whose task is to realize control according to instruction given in natural language. Note that such situation is not so unusual; remember, e.g., instruction given by teacher in car-driving school.
Using the software systems LFLCSim and LFLC2000 we developed several tens of simulations of control of various kinds of processes. The fuzzy PI, PD, and PID controllers have been implemented in them. A real application of linguistically oriented fuzzy control is described in [1].
Simple PI control
FControl-OP12.swf
Similar demonstration shows control of inverted pendulum.
FControl-Pendulum.swf
Universal PI control
There exists general linguistic description for PI fuzzy control of wide range of processes.
| No. | E | dE | dU | No. | E | dE | dU |
| 1. | Ze | Ze | Ze | 19. | -Me | +Sm | -ExSm |
| 2. | +Bi | +NoZe | +Bi | 20. | +Sm | +NoSm | +Me |
| 3. | -Bi | -NoZe | -Bi | 21. | -Sm | -NoSm | -Me |
| 4. | +Bi | -NoSm | Ze | 22. | +Sm | -NoSm | -VeSm |
| 5. | -Bi | +NoSm | Ze | 23. | -Sm | +NoSm | +VeSm |
| 6. | +Bi | RoZe | +Me | 24. | +Sm | RoZe | +VeSm |
| 7. | -Bi | RoZe | -Me | 25. | -Sm | RoZe | -VeSm |
| 8. | +Bi | -Sm | +Me | 26. | +Sm | +Sm | +Sm |
| 9. | -Bi | +Sm | -Me | 27. | +Sm | -Sm | +ExSm |
| 10. | +Me | +NoSm | +Bi | 28. | -Sm | +Sm | -ExSm |
| 11. | -Me | -NoSm | -Bi | 29. | -Sm | -Sm | -Sm |
| 12. | +Me | -NoSm | -VeSm | 30. | +VeSm | -Sm | -Sm |
| 13. | -Me | +NoSm | +VeSm | 31. | -VeSm | +Sm | +Sm |
| 14. | +Me | RoZe | +Me | 32. | +VeSm | -VeSm | Ze |
| 15. | -Me | RoZe | -Me | 33. | -VeSm | +VeSm | Ze |
| 16. | +Me | +Sm | +Me | 34. | RoZe | +VeSm | +ExSm |
| 17. | -Me | -Sm | -Me | 35. | RoZe | -VeSm | -ExSm |
| 18. | +Me | -Sm | +ExSm |
|
|
|
|
The following set of demonstrations shows the use of this linguistic description in the control of 6 different processes. The description is in all cases the same, only linguistic context (scaling) of the variables had to be set.
FControl-PIDipl1.swf
FControl-PIDipl2.swf
FControl-PIDipl3.swf
FControl-PIDipl4.swf
FControl-PIDipl5.swf
FControl-PIDiplUnstable.swf
Learning fuzzy control
REFERENCES:
[1] NOVÁK, V. Foundations of fuzzy modeling (Základy fuzzy modelování). Praha: BEN-technická literatura, 2000. 166 pp. ISBN 80-7300-009-1 (in Czech).
Plenary talk at 15th Carpathian International Control Conference, Velke Karlovice, Czech Republic 2014.Magnetic levitation using PID controller
Magnetic levitation using LFLC Controller
