联系我们： 手动添加方式: 微信>添加朋友>企业微信联系人>13262280223 或者 QQ: 1483266981

PART A- FUZZY DECISION-MAKING
Consider the process relating to muscle relaxation, as seen in your Laboratory Session
PART A, represented by the following Equations [PLEASE NOTE THE NEW VALUE
RELATING TO THE PREDOMINANT TIME-CONSTANT- CHANGE IT IN YOUR
MATLAB-SIMULINK MODEL]:
(A.1)
(A.2)
where Y is the overall output of the system (muscle relaxation), U is the input to the system
(amount of drug infused).
The main objective is to design a closed-loop control strategy which should maintain a steady
level of muscle relaxation (output ‘Y’) by manipulating the level of drug infused (input ‘U’),
given a reference target of muscle relaxation (‘Ref’)- see Figure A1- Use a reference target
of 0.8 all throughout and a simulation time of 300 minutes:
a) Simulate this process in closed-loop, via SIMULINK, just by using a simple negative
feedback loop and show/explain how difficult it is to maintain an accurate level of
muscle relaxation (output) in response to a target relaxation value.
[3 MARKS]
b) Design a Fuzzy PI-type controller, as shown in Figure A1, via a fuzzy rule-base which
should include 25 fuzzy rules with Gaussian Membership Functions.
[The candidate is expected to use: 1. the fuzzy 3D surface to tune the rules; 2.
His/her knowledge of tuning the PID tuning factors, all in order to obtain the best
possible outcome for the output response in terms of minimum overshoot, fast
rise-time and fast settling-time; include simulations with disturbances].
[16 MARKS]
c) Transform the fuzzy rule-base designed in A)b) into a Fuzzy PD-type controller which
should also include 25 fuzzy rules with Gaussian Membership Functions.
[The candidate is expected to use: 1. the fuzzy 3D surface to tune the rules; 2.
His/her knowledge of tuning the PID tuning factors, all in order to obtain the best
possible outcome for the output response in terms of minimum steady-state
error, minimum overshoot, fast rise-time and fast settling-time; include
simulations with disturbances]PART A- FUZZY DECISION-MAKING
Consider the process relating to muscle relaxation, as seen in your Laboratory Session
PART A, represented by the following Equations [PLEASE NOTE THE NEW VALUE
RELATING TO THE PREDOMINANT TIME-CONSTANT- CHANGE IT IN YOUR
MATLAB-SIMULINK MODEL]:
(A.1)
(A.2)
where Y is the overall output of the system (muscle relaxation), U is the input to the system
(amount of drug infused).
The main objective is to design a closed-loop control strategy which should maintain a steady
level of muscle relaxation (output ‘Y’) by manipulating the level of drug infused (input ‘U’),
given a reference target of muscle relaxation (‘Ref’)- see Figure A1- Use a reference target
of 0.8 all throughout and a simulation time of 300 minutes:
a) Simulate this process in closed-loop, via SIMULINK, just by using a simple negative
feedback loop and show/explain how difficult it is to maintain an accurate level of
muscle relaxation (output) in response to a target relaxation value.
[3 MARKS]
b) Design a Fuzzy PI-type controller, as shown in Figure A1, via a fuzzy rule-base which
should include 25 fuzzy rules with Gaussian Membership Functions.
[The candidate is expected to use: 1. the fuzzy 3D surface to tune the rules; 2.
His/her knowledge of tuning the PID tuning factors, all in order to obtain the best
possible outcome for the output response in terms of minimum overshoot, fast
rise-time and fast settling-time; include simulations with disturbances].
[16 MARKS]
c) Transform the fuzzy rule-base designed in A)b) into a Fuzzy PD-type controller which
should also include 25 fuzzy rules with Gaussian Membership Functions.
[The candidate is expected to use: 1. the fuzzy 3D surface to tune the rules; 2.
His/her knowledge of tuning the PID tuning factors, all in order to obtain the best
possible outcome for the output response in terms of minimum steady-state
error, minimum overshoot, fast rise-time and fast settling-time; include
simulations with disturbances]
d) Using the closed-loop data from A)b) derive an ANFIS based controller for the process,
which is described by Equations (A.1) and (A.2), to achieve a similar control
performance as that in b).
[10 MARKS]
e) Compare the controllers in A)b), A)c), and A)d) in terms of: flexibility in the structure
(controller type) and performance (accuracy). For the latter you can rely on one or
more performance indices, e.g. Mean Absolute Error (MAE), Mean Square Error
(MSE), and Root-Mean Square Error (RMSE).
[5 MARKS]

PART B- FUZZY PREDICTIVE MODELLING
On Black-Board, you will find one (1) file named “acs323assignmentdata.mat” (in the folder
“Module Assignment”) relating to industrial data. This data set should reflect a system with
four (4) inputs (Input 1, …, Input 4) and one (1) output. The minimum and maximum values
for all inputs and outputs in the provided data can be found using the “min” and “max” MATLAB
commands.
Upload this file onto your local drive which you will subsequently use to carry-out tasks B)a)-
B)e) in MATLAB.
Once you have uploaded this file in MATLAB and double-clicked on it, it will collapse into two
(2) files: one file, “acs323assignmentdata”, contains the actual quantitative data, and the
second file, “explanation”, provides details on the names for each of the five (5) data features.
a) Use the ANFIS tool in MATLAB to obtain a fuzzy TSK-type model, with 3 membership
functions for each input. The fitness of the model should be assessed with the use of
a quantitative index (or indices) such as the RMSE, MSE or MAE to establish the
validity of the model. You can use one performance index or more than one all
throughout.
[The student will partition the data accordingly, select the most appropriate type
of fuzzy MFs, output function and the number of learning epochs which will lead
to the best outcome in the least-square sense].
[18 MARKS]
b) Using the model derived in B)a) find the values of the output for the following input
vectors:
Input 1 = 0.15; Input 2 = 0.22; Input 3 = 1; Input 4 = 0.011;
Input 1 = 0.06; Input 2 = 0.28; Input 3 = 0.4; Input 4 = 0.012;
[4 MARKS]
c) Extend the fuzzy modelling exercise conducted in B)a) to include 4 membership
functions for each input, then 5 membership functions for each input. Here also, the
fitness of the models should be assessed with the use of a quantitative index (or
indices) such as the RMSE, MSE or MAE to establish the validity of the models.
[The student will partition the data accordingly, select the most appropriate type
of fuzzy MFs, output function and the number of learning epochs which will lead
to the best outcome].
[18 MARKS]
d) Using the models derived in B)c) find the values of the output for the following input
vectors:
Input 1 = 0.15; Input 2 = 0.22; Input 3 = 1; Input 4 = 0.011;
Input 1 = 0.06; Input 2 = 0.28; Input 3 = 0.4; Input 4 = 0.012;
[4 MARKS]
e) Compare the models derived in B)a) and B)c) and draw your own conclusions with
respect to model accuracy and generalisation properties as far as: 1. Data partitioning
between training and testing; 2. The number of fuzzy MFs, are concerned.
[6 MARKS]