Academia.eduAcademia.edu

Figure from:A survey on industrial applications of fuzzy control

See full PDFdownloadDownload figure
Fig. 5. Logical scheme of an adaptive fuzzy system for cruise control. EAV is the error average value, UAV is the average value of the throttle opening, Ve;; is the maximum value of the error [197].

Figure 5 Logical scheme of an adaptive fuzzy system for cruise control. EAV is the error average value, UAV is the average value of the throttle opening, Ve;; is the maximum value of the error [197].

Related Figures (8)

Fig. 1. Basic fuzzy control system structure. The block diagram of principle (considered as classical in the literature) of an FCS is presented in Fig. 1. The FCS is considered as a single input system with respect to the reference input rand asa
Fig. 1. Basic fuzzy control system structure. The block diagram of principle (considered as classical in the literature) of an FCS is presented in Fig. 1. The FCS is considered as a single input system with respect to the reference input rand asa
Fig. 2. Set-point filter 2-DOF control system structure (a), feedforward 2-DOF control system structure (b), feedback 2-DOF control system structure (c), component-separated 2-DOF PI control system structure (d), definition of load disturbance input scenarios (e).
Fig. 2. Set-point filter 2-DOF control system structure (a), feedforward 2-DOF control system structure (b), feedback 2-DOF control system structure (c), component-separated 2-DOF PI control system structure (d), definition of load disturbance input scenarios (e).
Fig. 3. An adaptive fuzzy controller with a supervisory system. There are many formations for the FC in FCSs similarly to the different control schemes focused on PI controllers presented in Fig. 2. An adaptive FC has one extra component, a supervisory module, as shown in Fig. 3. The supervisory module has
Fig. 3. An adaptive fuzzy controller with a supervisory system. There are many formations for the FC in FCSs similarly to the different control schemes focused on PI controllers presented in Fig. 2. An adaptive FC has one extra component, a supervisory module, as shown in Fig. 3. The supervisory module has
Fig. 4. Changing the side slopes of a group of membership functions. The modal values (centres) stay at the same position.
Fig. 4. Changing the side slopes of a group of membership functions. The modal values (centres) stay at the same position.
Parking garages and the prediction quality of the adapted fuzzy system.
Parking garages and the prediction quality of the adapted fuzzy system.
Fig. 6. Repositioning fuzzy sets such that they fit to clusters of input data.
Fig. 6. Repositioning fuzzy sets such that they fit to clusters of input data.
Fig. 7. Structure of fuzzy parking system.
Fig. 7. Structure of fuzzy parking system.
Fig. 8. A neural network translation from a fuzzy rule base. SOS eee aye ese | GRR Bey Ret This method. is, , for example, used in a system with many sensors in a car that had to diagnose the kind of road and traffic conditions the car was driving in [204]. There was an initial rule base describing the logical relation between the sensors and the driving situation. Test data were generated by driving the car with a video camera several days in all kinds of situations. The test data were categorized to several driving situations. After translating the rule base to a neural network all rule weights, the positions of the membership functions and the link weights were adapted which resulted in a much improved fuzzy rule base. The customer demanded a rule base that could be checked by hand afterwards
Fig. 8. A neural network translation from a fuzzy rule base. SOS eee aye ese | GRR Bey Ret This method. is, , for example, used in a system with many sensors in a car that had to diagnose the kind of road and traffic conditions the car was driving in [204]. There was an initial rule base describing the logical relation between the sensors and the driving situation. Test data were generated by driving the car with a video camera several days in all kinds of situations. The test data were categorized to several driving situations. After translating the rule base to a neural network all rule weights, the positions of the membership functions and the link weights were adapted which resulted in a much improved fuzzy rule base. The customer demanded a rule base that could be checked by hand afterwards