INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
Design and Validation of a Mamdani-Type Fuzzy Inference System  
for Dynamic Indoor Climate Balancing  
Engr. Prince Jaminn A. Soberano; Engr. Charmaine L. Robles, PECE; Dr. Ma. Magdalena V. Gatdula  
Graduate School, Bulacan State University  
DOI: https://dx.doi.org/10.51244/IJRSI.2025.12110039  
Received: 18 November 2025; Accepted: 27 November 2025; Published: 04 November 2025  
ABSTRACT  
The goal of this research is the design, simulation, and validation of a stable and energy-efficient Mamdani-type-  
1 FLC that controls an indoor climate balancing system by overcoming the drawbacks of conventional linear  
control in handling the intrinsic nonlinearity and complexity of the system. The main objective will be to  
dynamically control crucial climate parameters such as Fan Speed and Cooling Rate based on crisp input values  
of Temperature in the range [8 44] and Relative Humidity in the range [0 90]. The operational intelligence of  
the FLC relies on a comprehensive fuzzy rule base of thirty-five (35) IF-THEN rules that connect seven fuzzy  
sets for temperature and five for humidity to their corresponding output actions. The simulation also highlights  
the capability of the FLC to smoothly offer nonlinear control transitions from minimum to maximum effort, thus  
avoiding abrupt on/off behavior that wastes energy. This research validates the FLC as an effective, feasible,  
and energy-efficient control solution, laying a very firm foundation for further research.  
Keywords— Fuzzy Logic Controller, Indoor Climate Balancing, Mamdani Inference, HVAC System, Fuzzy  
Rule Base  
INTRODUCTION  
Indoor Climate Balancing Systems, consisting of HVAC (heating, ventilating, and air conditioning) systems and  
RAC (refrigeration and air conditioning) systems, are a significant area for management since they account for  
a large portion of building energy consumption globally [1-3]. Energy consumption is critical for any interior  
climate balancing system; therefore, the goal should be to reduce energy consumption while maintaining  
appropriate and acceptable indoor temperatures for the occupants [1-2].  
Conventional control practices adopted for HVAC systems, such as ON/OFF or Proportional-Integral-Derivative  
(PID) controllers, often struggle to capture the complex nature of HVAC equipment, including nonlinear system  
dynamics, parameter uncertainties, and time-varying characteristics [1-2]. Optimal climate control demands the  
concurrent control of Temperature and Relative Humidity [4-5]. Linear control strategies underperform because  
they are insufficient for controlling the nonlinear features of air-conditioning systems, which behave as multiple-  
input multiple-output (MIMO) systems with closely linked parameters [2].  
Among the intelligent control techniques, Fuzzy Logic Controllers arise as one of the most advanced and  
efficient alternatives to surpass the limitations of conventional control systems [6]. Fuzzy logic systems imitate  
the way humans think by implementing linguistic principles, which make it possible to handle nonlinear and  
mathematically complicated systems without the need for a detailed mathematical representation [4, 6-7].  
Therefore, the goal of this study is to take advantage of the merits of FLC in developing a multi-output fuzzy  
model that would dynamically balance inputs like temperature and humidity with control outputs like fan speed  
and cooling rate, thus reducing energy consumption and enhancing the efficiency of the system.  
In most parts of the world, HVAC systems are one of the largest electrical power consumers, utilizing about  
40% to 60% of a building's total energy consumption [1]. The operation of HVAC systems continues to be more  
challenging due to the complexity and nonlinearity of the system and because it is a multi-input, multi-output  
device with interconnected characteristics that cannot be optimally regulated by standard linear control methods  
Page 427  
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
such as PID [2, 8]. From a performance point of view, conventional controllers often control temperature and  
humidity, the two most important comfort variables, leading to poor performance, energy losses, and wear of  
actuators through oscillation or unexpected changes [2, 9]. Therefore, FLCs are applied in intelligent control  
systems, which can handle complexity and uncertainty without explicit mathematical modeling of the system  
[10-11].  
The intelligent control system will act to adjust the important climate parameters, such as Fan Speed and Cooling  
Rate, based on the assessment of inputs like Temperature and Relative Humidity, which would help to overcome  
inadequacies related to handling the system's inherent nonlinearity and complexity by conventional linear  
controllers.  
The following specific objectives will be useful in achieving this overall goal:  
1. To establish the architecture of the FLC including the selection of Mamdani-type inference for efficient  
operation in HVAC/RAC applications.  
2. To characterize the input variables, Temperature and Humidity, with appropriate ranges and linguistic labels,  
and define the resulting output control actions, Fan Speed and Cooling Rate, along with their respective  
ranges and linguistic labels.  
3. To map crisp input values, such as Temperature and Humidity, to fuzzy linguistic sets ("Cold," "High")  
through appropriate membership functions in the fuzzification stage of the control system.  
4. To construct a comprehensive set of IF-THEN rules that replicate human reasoning and govern the necessary  
control actions for maintaining an optimal indoor climate based on the coupled Temperature and Humidity  
inputs.  
5. To implement and simulate the designed FLC model using MATLAB and test its performance across various  
indoor conditions, ensuring the FLC provides stable, adaptive, and energy efficient control actions.  
METHODOLOGY  
The research methodology on designing and implementing the Indoor Climate Balancing System using a  
Mamdani-type Fuzzy Logic Controller includes system modeling, design steps specific to the FLC, and  
performance validation as illustrated in Fig. 1.  
Fig. 1 System Design Diagram  
A fuzzified controller is a type of tool that takes real world measurements, converts them into fuzzy values,  
refines them by rule-based inference machine, and finally transforms fuzzy output values into definite control  
signals [7]. The FLC serves as the backbone of the system and utilizes the Mamdani-type inference, one of the  
most widely applied in HVAC/RAC applications [7]. The sequence in fuzzy logic starts with inputting numerical  
values (Crisp Inputs) from the sensor, e.g., indoor temperature and humidity. In contrast to crisp inputs, the  
Page 428  
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
Fuzzy Inference System uses fuzzy-like terms or can be described as subjective linguistic terms. For example, a  
temperature reading of 30°C is converted into a degree of membership for the linguistic sets "Warm" or "Hot"  
using membership functions (MFs).  
This fuzzified data is sent into the Inference Engine, which uses a set of IF-THEN rules (for example, IF  
Temperature is "Warm" AND Humidity is "High," THEN Cooling Rate is "Medium"). To select the fuzzy  
output, the Inference Engine follows the procedure of human reasoning, usually using Mamdani-Type (Max-  
Min) inference 7. The last process is the Defuzzification, which reconverts the fuzzy output into a practical, crisp  
value. Outputs, which are a precise signal, are then assigned to Fan Speed and Cooling Rate to maintain the  
desired environmental conditions.  
Table I Table of Variables  
Parameter  
Type  
Range  
Linguistic Labels  
Temperature  
Input  
[8 44]  
Cold, Cool, Normal, Warm, Hot,  
Very Hot, Extremely Hot  
Humidity  
Input  
[0 90]  
[0 3]  
Very Low, Low, Normal, High,  
Very High  
Cooling Rate  
Output  
Extremely Low, Very Low, Low,  
Normal, High Very High, Extremely  
High  
Fan Speed  
Output  
[0 100]  
OFF, Low, Medium, High, Max  
Table 1 shows that the FLC for the Indoor Climate Balancing System is designed to use two different input  
control variables: Temperature and Humidity. A temperature input ranges between 8 and 44 and is described by  
seven linguistic labels: Cold, Cool, Normal, Warm, Hot, Very Hot, and Extremely Hot. The second input,  
Humidity, ranges between 0 and 90 and is described by five linguistic labels: Very Low, Low, Normal, High,  
and Very High. It is important to mention that these numeric ranges of Temperature and Humidity were taken  
from the methodology in reference [6, 12] to make the system applicable within the context of indoor climate  
control.  
These fuzzified inputs, processed by the Fuzzy Inference Engine using the rule base, determine the system's two  
outputs: Fan Speed and Cooling Rate. The Fan Speed has a numeric range of 0 to 100 and uses five linguistic  
labels: OFF, Low, Medium, High, and Max. On the other hand, the Cooling Rate is highly limited in its range  
of 0 to 3, and is controlled by seven linguistic labels: Extremely Low, Very Low, Low, Normal, High, Very  
High, and Extremely High. In general, this structure enables the FLC to convert the fuzzy perception of the  
indoor climate-through the defined input sets-into precise, granular crisp commands in both the intensity of  
cooling and the velocity of airflow, maintaining the desired indoor climate condition.  
The operational intelligence of the FLC is embedded within its comprehensive Fuzzy Rule Base, comprising  
thirty-five (35) distinct IF-THEN rules shown in Figure 2. These rules are developed to try and emulate human  
decision-making and expert knowledge by using the logical AND operator that links the two input linguistic  
variables: Temperature (e.g., Cold, Normal, Hot) and Humidity (e.g., Very Low, Normal, High), while each of  
the rules then prescribes the control action for the two output linguistic variables: Fan Speed (e.g., OFF, Low,  
Max) and Cooling Rate (e.g., Extremely Low, Normal, Extremely High). For example, a rule might be: "IF  
Temperature is Hot AND Humidity is High THEN Fan Speed is Max, Cooling Rate is Very High." Such a  
complicated network of rules provides an option for the FLC to estimate precisely which mix of fan speed and  
cooling intensity is proper under all possible combinations of temperature and humidity conditions for adaptive,  
stable, and energy-efficient climate control. The systematic enumeration of these rules covers the entire input  
space, guaranteeing an efficient and responsive system performance.  
Page 429  
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
Fig. 2 IF-THEN Rules  
RESULTS AND DISCUSSION  
The implementation of the research for indoor climate control, details the design and application of a Mamdani  
Type-1 Fuzzy Logic Control system. The goal of the implemented intelligent system is to analyze the automation  
of cooling systems by applying fuzzy logic rules based on temperature and relative humidity parameters to  
control the surrounding air and power consumption for indoor environment. The system automatically controls  
the cooling devices when the temperature varies between 0 °C to 50 °C and relative humidity varies between  
0% to 90% which was adapted in the previous researches on Fuzzy logic systems [6, 12].  
The core intelligence is managed by the FLC, which was designed and simulated using the MATLAB platform.  
The model utilized the Mamdani’s method Type-1. The Mamdani FIS was chosen because it includes an output  
membership function, unlike the Sugeno FIS. The schematic of the fuzzy system is composed of three major  
components (Fuzzification, Fuzzy Inference System, and Defuzzification). For the fuzzification, the crisp inputs  
(Temperature and Humidity) are mapped to fuzzy sets. Seven (7) triangular fuzzy sets are selected for the input  
temperature [12], and five (5) triangular linguistic fuzzy sets are selected for relative humidity [6]. While the  
Fuzzy Inference System will produce the output according to the system inputs, utilizing thirty-five (35) IF-  
THEN rules that are constructed with the AND operation. These rules convert the degrees of membership for  
the inputs into an output fuzzy set. Lastly, the Defuzzification generates specific fuzzy output values resulting  
from the inference system and converts them into crisp values to manage the cooling system mode setting.  
Table II Simulation Results  
Actual  
Output  
[FS CR]  
Expected  
Output  
Test No.  
Input 1  
Input 2  
System Response / Observation  
The system responds in a stable,  
but forceful manner. The Fan  
Speed (medium range) with a  
non-zero Cooling Rate (Normal  
range) is adaptively neutral. The  
model views the state as mildly  
cool but not excessively dry and  
on the threshold of its comfort  
zone.  
FS will be at low  
range and CR  
will be Low  
1
15  
30  
[50 1.5]  
The system response is stable and  
subtle. This is low sensitivity  
because even though the  
FS will be at low  
range and CR  
will be  
2
17  
61  
[95.7 2.89]  
conditions are not exactly  
Page 430  
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
Extremely Low  
Normal, the system resists  
forceful action, instead of  
accelerating the fan and creating a  
significant amount of cooling.  
This adaptive stability maintains  
the system operating with very  
little energy and not over-  
correcting for the minor humidity  
issue near the optimal  
temperature of the room.  
The system reacts rapidly and  
adaptively with medium-to-high  
outputs. This reaction  
demonstrates the system is  
sensitive to discomfort,  
FS will be at  
medium range  
and CR will be  
normal  
automatically engaging a medium  
range fan speed and using a  
significant amount of cooling. It  
is stable because the outputs are a  
balanced trade-off across several  
active rules, indicating it can  
adjust to situations between its  
primary comfort levels without  
merely maxing out.  
3
26  
45  
[60.4 1.87]  
The system responds fully and  
strongly with high efficiency.  
This is evidence of high  
sensitivity to extreme heat,  
instantly demanding a lot of  
cooling and high speed of the fan.  
But the Cooling Rate is slightly  
dragged below the maximum  
since the humidity is low,  
FS will be at  
max range and  
CR will be very  
high  
4
38  
28  
[79.1 2.37]  
indicating the system is adaptive  
and able to slightly modify its  
effort according to dryness, while  
being fully stable in its effort to  
lower the temperature.  
The system reacts rapidly and  
forcefully with the greatest  
possible outputs. The reaction is  
most sensitive to the inputs,  
calling for an extremely high  
effort to cool and dry out at once,  
which demonstrates its  
FS will be Max  
and CR will be  
Extremely High  
5
40  
79  
[95.7 2.89]  
immensely adaptive to extreme  
conditions and fully stable in  
providing the most reasonable  
action.  
Page 431  
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
The simulation effectively demonstrates the Mamdani Type-1 FLC system's ability to provide a stable and  
adaptable indoor environment control. There are thirty-five (35) IF-THEN rules with seven (7) fuzzy sets for  
temperature and five (5) for humidity which are the primary process utilized in computing the FS and CR  
outputs. The outputs show that the system is highly adaptive and can perform subtle decisions rather than a  
conventional on/off switch. For example, under near ideal conditions (i.e., 15 °C and 30 % humidity), the non-  
zero Cooling Rate and low fan speed exemplify adaptive neutrality rather than the forcing of action that wastes  
energy. At the highest rating for temperature and humidity, for instance, 40°C and 79% relative humidity  
respectively, the system is most sensitive with both outputs controlled as near to their maximum, FS and CR at  
95.7 and 2.89 respectively, as practical in attempt to make quick comfort correction feasible. This smooth  
transition from zero effort to a maximum output seeks to show the FLC model as a workable and practical  
method for the automation of climate balancing systems.  
The results of the simulation can be visually represented by the Figure 3 and Figure 4 Surface Viewer, which  
graphically illustrates the relationship between the two input variables-the Temperature and Humidity-and the  
two dynamically controlled output variables-Fan Speed and Cooling Rate. These surfaces confirm the  
effectiveness of the Mamdani Type 1 FLC in providing smooth, stable, and adaptive control across the full range  
of possible environmental conditions.  
Fig. 3 Cooling Rate Surface  
Fig. 4 Fan Speed Surface  
Fig. 3 presents a clear, logical control strategy for CR surface, whose output (Z-axis) increases smoothly as the  
inputs (X and Y axes) move away from the ideal comfort zone. Similarly, Figure 4 shows a highly correlated  
output, which is expected since the fan is supposed to distribute the cooled and dried air. Close to the ideal  
comfort zone, the FS output assumes a low or zero setting, confirming the low sensitivity and subtlety in the  
system response. For increasing temperature and humidity inputs, the Fan Speed smoothly ramps up to its  
maximum range, ensuring the immediate, fast dissemination of cooling required for correcting extreme  
conditions to retain stability and realize the most reasonable action. The continuous, nonlinear shape of both the  
control surfaces confirms that the FLC provides a smooth transition from zero effort to maximum output,  
Page 432  
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
avoiding the suboptimal performance and actuator wear caused by the oscillations typical of conventional on/off  
controllers.  
Similarly, from Figure 4, the output is highly correlated, as would be expected from a fan that is supposed to  
disseminate the cooled and dried air. For a temperature and humidity input close to the ideal comfort zone, the  
FS output assumes a low or zero setting, confirming that the response of the system is of low sensitivity and  
subtlety. With increasing temperature and humidity inputs, the Fan Speed smoothly ramps up to its maximum  
range in order to ensure the immediate, fast dissemination of cooling required for the correction of extreme  
conditions to retain stability and realize the most reasonable action. The continuous, nonlinear shape of both  
control surfaces confirms that the provided FLC offers a smooth transition from zero effort up to a maximum  
output, avoiding suboptimal performance and wear of actuators due to the oscillations typical of conventional  
on/off controllers.  
The continuity of both 3D surfaces and their non-linear behavior prove that FLC manages a smooth transition  
from zero effort to maximum output. The FLC also avoids the abrupt, energy-wasting on/off behavior typical  
for conventional controllers. Based on the graphical validation, the Mamdani FLC would be a feasible, practical,  
and robust solution to automate climate-balancing systems while optimizing energy efficiency.  
CONCLUSIONS  
Design and simulation proved that the Mamdani Type 1 FLC successfully achieved the objective of stable and  
energy-efficient control in Indoor Climate Balancing Systems. Replacing the complex, nonlinear mathematical  
modeling of MIMO systems, such as HVAC, it embodies a computationally simple, robust control structure  
based on human-like linguistic reasoning. With a total of thirty-five IF-THEN rules based on seven fuzzy sets  
for Temperature and five for Humidity, the prominent adaptivity of the FLC was demonstrated with smooth  
transitions of the outputs-Fan Speed and Cooling Rate-from subtle, non-zero actions near the ideal comfort  
conditions to a maximum required effort in extreme environmental states. This finally justifies the FLC model  
presented here as feasible and practical for automation of climate-balancing systems, which is continuously  
optimized for energy efficiency compared to conventional methods.  
Although the FLC model was robust in simulation, one of the major limitations is that these findings are based  
solely on the Mamdani inference method. No comparative study was conducted against other prominent fuzzy  
models; thus, absolute performances and computational efficiency relative to alternatives remain unidentified.  
It is highly recommended to conduct a comparative study of the present FLC against other fuzzy inference  
models, specifically the Sugeno model and ANFIS models, in future research. Such a comparative study will be  
imperative for full validation of the findings and to identify, beyond any doubt, the most effective and  
computationally superior model for implementation in a real-world indoor climate balancing system.  
ACKNOWLEDGEMENT  
This research was completed as a self-funded academic project, and as such, the authors sincerely acknowledge  
that no external funding institution or organization provided financial or material support for the design,  
simulation, or write-up of the Fuzzy Logic Controller for the Indoor Climate Balancing System. The successful  
completion of this study is entirely attributed to the personal time, dedication, and technical resources of the  
research group.  
REFERENCES  
1. Reddy, “Application of Fuzzy Logic Controllers in HVAC Systems for Energy Optimization,”  
International Journal of Research in Modern Engineering & Emerging Technology (IJRMEET), Jul. 02,  
2021.https://ijrmeet.org/application-of-fuzzy-logic-controllers-in-hvac-systems-for-energy-  
optimization/  
2. F. Behrooz, N. Mariun, M. Marhaban, M. Mohd Radzi, and A. Ramli, “Review of Control Techniques  
for HVAC Systems—Nonlinearity Approaches Based on Fuzzy Cognitive Maps,” Energies, vol. 11, no.  
3, p. 495, Feb. 2018, doi: https://doi.org/10.3390/en11030495.  
Page 433  
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)  
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XI November 2025  
3. M. Wadia and K. Beragi, “Review on Matlab Fuzzy Logic Based and Predictive (Radial Basis Function,  
Rbf) Hvac Controllers,” International Journal of Scientific Research & Engineering Trends, vol. 9, no.  
6,  
pp.  
2395–566,  
2023,  
Accessed:  
Oct.  
18,  
2025.  
[Online].  
Available:  
https://ijsret.com/wpcontent/uploads/2023/11/IJSRET_V9_issue6_443.pdf  
4. K. A. Akpado, P. N. Nwankwo, D. A. Onwuzulike, and M. N. Orji, “A HYBRID APPROACH FOR  
AIR CONDITIONING CONTROL SYSTEM WITH FUZZY LOGIC CONTROLLER,” International  
Journal of Engineering and Applied Sciences (IJEAS), vol. 5, no. 8, Aug. 2018, doi:  
https://doi.org/10.31873/ijeas.5.8.01.  
5. Yadav and A. Goel, “Comparative Analysis of HVAC using PID, Fuzzy and ANFIS Technique,”  
International Journal of Research in Advent Technology, vol. 6, no. 7, 2018, Accessed: Oct. 18, 2025.  
[Online]. Available: https://ijrat.org/downloads/Vol-6/july-2018/paper%20ID-672018102.pdf  
6. H. N. Y. Al-Talb, S. N. M. Al-Faydi, T. A. Fathi, and M. A. S. Al-Adwany, “A Fuzzy Logic IoT- Based  
Temperature and Humidity Control System for Smart Building,” International Journal of Computing and  
Digital Systems, vol. 13, no. 1, pp. 139–147, Jan. 2023, doi: https://doi.org/10.12785/ijcds/130111.  
7. J. M. Belman-Flores, D. A. Rodríguez-Valderrama, S. Ledesma, J. J. García-Pabón, D. Hernández, and  
D. M. Pardo-Cely, “A Review on Applications of Fuzzy Logic Control for Refrigeration Systems,”  
Applied Sciences, vol. 12, no. 3, p. 1302, Jan. 2022, doi: https://doi.org/10.3390/app12031302.  
8. Z. Nomani and P. Tripathi, “Design and Implementation of Fuzzy and ANFIS Controller in HVAC  
System for Better Energy Management: A Survey,” International Journal of Engineering, vol. 8, no. 6,  
p. 264798, Jun. 2018.  
9. M. A. Abuhussain, Badr Saad Alotaibi, Muhammad Saidu Aliero, M. Asif, M. Alshenaifi, and Yakubu  
Aminu Dodo, “Adaptive HVAC System Based on Fuzzy Controller Approach,” Applied sciences, vol.  
13, no. 20, pp. 11354–11354, Oct. 2023, doi: https://doi.org/10.3390/app132011354.  
10. Chojecki, A. Ambroziak, and P. Borkowski, “Fuzzy Controllers Instead of Classical PIDs in HVAC  
Equipment: Dusting Off a Well-Known Technology and Today’s Implementation for Better Energy  
Efficiency and User Comfort,” Energies, vol. 16, no. 7, pp. 2967–2967, Mar. 2023, doi:  
https://doi.org/10.3390/en16072967.  
11. P. Singhala, D. N. Shah, and B. Patel, “Temperature Control using Fuzzy Logic,” International Journal  
of Instrumentation and Control Systems, vol. 4, no. 1, pp. 1–10, Jan. 2014, doi:  
https://doi.org/10.5121/ijics.2014.4101.  
12. T. Das and Y. Das, “Design of A Room Temperature And Humidity Controller Using Fuzzy Logic,”  
American Journal of Engineering Research (AJER), vol. 02, no. 11, pp. 86–97, 2013, Accessed: Nov.  
08, 2025. [Online]. Available: https://www.ajer.org/papers/v2(11)/J02118697.pdf  
Page 434