The paper investigates the use of fuzzy data converters based on intelligent systems of fuzzy inference in the Internet of Things systems. An approach has been developed for converting fuzzy input parameters read from sensors into a PWM signal based on the Takagi-Sugeno fuzzy algorithm. Based on the suggested method an IoT-based smart fan was developed to demonstrate the characteristics of the proposed method for intelligent control of Internet of Things devices that support a PWM signal. Finally, the findings of the experimental data were utilized to simulate the performance of the fuzzy model, using three parameters: temperature, relative humidity, and carbon dioxide concentrations (CO2) vis-à-vis the PWM signal output. The proposed method shows the simplicity of training a smart fan control system and makes the possibility of efficient energy consumption. Control systems, fuzzy inference, Internet of Things, fuzzy PWM-controller, Smart fan.
2021 Copyright for this paper by its authors. 2010 and 2050 [7]. Therefore, the development of IoT-based platforms in the field of smart ventilation needs to be improved in terms of resource consumption. We must develop new techniques for improving the energy efficiency of ventilation systems to achieve the reduction in overall building energy consumption and as a result decrease operating electricity costs.
IoT smart fan systems still rely on the traditional methods of control strategies such as cycling or on/off control, staging, modulation, or proportional control [8]. These may be quite obsolete, inaccurate, and inefficient for energy consumption in smart fan systems. The novel idea is to apply fuzzy logic in early IoT-based fans to assess the environmental factors like temperature, humidity, carbon dioxide concentrations, etc and convert them by Takagi-Sugeno [9] algorithm into Pulse width modulation (PWM) signal to control the speed of motors, heat output of heaters, and the other parameters in an energy-efficient and quieter manner [10]. Thus, the proposed smart inverter fan based on fuzzy logic will provide comfortable levels of cooling and optimized electricity consumption.
The main contribution of the paper is: To consider the Takagi-Sugeno algorithm for input data conversion in IoT-based fans with a fuzzy expert system. To propose the use of Takagi-Sugeno systems for converting input data from sensors into an output PWM signal for efficient energy consumption in smart ventilation systems. The experimental results demonstrate dependencies between input and output variables of a smart fan controller based on the fuzzy converter, which transforms quality indicators into PWM signal, which assume its high performance and cost-effectiveness.
The paper is structured as follows: Section 1 introduces the existing methods of control strategies in smart ventilation systems and describes the energy consumption problem. Section 2 introduces the fundamentals of fuzzy control systems and describes the Takagi-Sugeno algorithm. In Section 3, Takagi-Sugeno system for converting output data from sensors into an output PWM signal is proposed to provide efficient energy consumption in smart air ventilation systems. In Section 4, the experiment results are demonstrated. Finally, Section 5 concludes the research paper.
Intelligent fuzzy systems are actively used to solve a wide class of problems in many areas of industry and life (medicine, production, safety, management, etc.). Belonging to the class of intellectual, fuzzy systems are designed to solve highly specialized tasks of the creative direction. Such tasks may include decision-making systems, expert systems, artificial intelligence systems, testing, assessment, classification systems, etc. [11-16]. 2.1.
Control systems based on fuzzy converters of input data into output data are used in cases where the following features take place: The system operates with qualitative values and characteristics There is incomplete data about the modeled object and its environment The investigated object is extremely difficult to model and find ideal solutions There is a non-linear dependence of the input-output data Decision making by the system is based on the knowledge and experience of experts in a specific problem area
Figure 1 shows a general scheme of a fuzzy inference system, which is equally suitable for all fuzzy algorithms [11] (Takagi-Sugeno, Mamdani, Larsen ...). As can be seen in the figure, many input values = { : = ̅1̅̅,̅̅} are converted into a set of output conclusions = { : = ̅1̅̅,̅̅̅}, using an inference algorithm on a fuzzy knowledge base. The knowledge base, typical for any fuzzy inference algorithm, consists of blocks of fuzzy rules = { , = ̅1̅̅,̅̅}products of IF-THEN statements (1). : x y , (1) where x ⊆ X, y ⊆ Y, – input linguistic variable, – output linguistic variable respectively. 1 …
Depending on the fuzzy inference algorithm, fuzzy production rules can have a different structure (the consequent can be a fuzzy value or a precise number)
systems, the consequents of which are fuzzy values) of the rules are clear numbers)
Accumulation of sub-conclusions of the consequent of fuzzy rules (carried out only for those Defuzzification of output values (or a procedure similar to defuzzification, if the consequents Let us take a closer look at the fuzzy logical conclusion of Takagi-Sugeno [9], since for research in this work it is of the greatest interest in the context of fuzzy data converters in IoT systems. The main feature of Takagi-Sugeno systems is the ability to transform qualitative indicators (fuzzified values) into quantitative ones since the subconclusions of fuzzy rules consist of functional dependencies on a set of input data (primary, non-fuzzified) and generate precise numbers.
The Takagi-Sugeno fuzzy rule is: : x = 0 + ∑
, =1
input prerequisites. where j – system-generated output number, w – fuzzy rule number, n – number of input variables, – free coefficient. The consequent of a rule is essentially a weighted summation of non-fuzzified
An output conclusion for every (see Figure 1) according to the Takagi-Sugeno algorithm, it is the finding of the weighted average of rule subconclusions in a fuzzy knowledge base (3): = ∑ =1[⋀ =1 ( )]( 0 + ∑ =1
) ∑ =1[⋀ =1 ( )] = ∑ =1 (x)
∑ =1 (x) where ⋀ − is the operation of taking the minimum, (х) – is the membership function of the input value to a fuzzy term. (2) (3) 3. Control of the output PWM signal in IoT systems based on intelligent fuzzy converters (fuzzy PWM-controller)
PWM-signal [10, 17] (pulse width modulation) – one of the varieties of a digital signal and is designed to control the power (speed) on output devices based on periodic on / off voltage on the port (Figure 2). One of the tasks related to the operation of the microcontroller in IoT systems is to control the PWM signal, and often its value depends on the input parameters read from other sensors (temperature, smoke sensor, proximity, etc.).
The total power consumption at the PWM output is calculated as: = + + + , (4) where – the energy consumption during the time of transition from Low to High mode, – the energy consumption during the current supply time to maintain High mode, – the energy consumption during the time of transition from High to Low mode, during the time for current regeneration in order to maintain Low mode, – the energy consumption – total operating time of the PWM signal (0..255).
Thus, there is often a non-linear relationship between the input parameters and the PWM output signal in such systems. It should be noted that the read-out data from the sensors are very often qualitative indicators and require the introduction of fuzzy linguistic values for their further operation and classification.
The paper proposes the use of Takagi-Sugeno systems for converting input data from sensors into an output PWM signal. A characteristic feature of this approach is the ability to convert input data from several sensors, realizing a nonlinear input-output dependence (a model of a converter of quality data from sensors into a PWM signal is shown in Figure 3). This approach allows the implementation of intelligent IoT systems with high energy savings due to the use of PWM as a control signal.
255 ( ) 255 255 qualitative characteristics read from the sensors:
In Figure 3, the input parameters are represented by a vector of fuzzy (fuzzified) values, which are ̃ = {̃ , = ̅1̅̅,̅̅}, ̃ = { 1( ) + 2( ) + ⋯ + ( ) }, where + is a union operation, ̃ – vector of fuzzy input values.
Subconclusions of fuzzy rules (Takagi-Sugeno fuzzy rules), form the weighted values of the PWM output signal from 0 to 255, which then go through the procedure for finding the weighted average value, which, accordingly, will also lie in the range [0..255]. t (mc) [0..255] [0..255] [0..255] [0..255]
Weighted average ∑ =1 (x) ∑ =1 (x)
PWMvalue [0..255] 4. Smart fan based on the fuzzy PWM-controller for convert of quality indicators into PWM signal 4.1.
In this paper, a model of an intelligent fan based on fuzzy control has been developed using the proposed in this article method of fuzzy transformation of quality indicators into a PWM signal. The use of a fuzzy controller allows the ventilation system to work in different modes of operation, adjusting to the external environment, and also taking into account the individual requirements of the customer, the climate of a particular region, medical contraindications, etc. Based on a fuzzy knowledge base, this system makes it possible to formalize the customer's requirements in the form of an expert set of rules, makes it possible to make the operating modes more varied.
It is also worth mentioning that efficient energy consumption, sustainability, environment-friendly – terms that should be taken into consideration in the modern scientific world. The heating, ventilation, and air conditioning systems can be the largest energy consumers in the building. The different approaches to modeling these systems and providing them with additional controllers can change the situation for the better.
The most common methods to solve the above problems are the classic use of proportionalintegral-differential (PID) controllers and Computational Intelligence techniques [18].
The advantage of using intelligent control methods in ventilation systems over the classic ones is the ability to regulate the room temperature at partial load, minimize system steady-state error. In recent studies [19–20] there was an attempt to use fuzzy logic to model the cooling process of ventilation systems. However, existing systems do not take into account that the resistance of transistors and resistors also leads to additional energy loss, because they burn part of it as a heat. Such systems are constantly running at full speed, making a lot of noise and consuming a lot of energy. The solution may be to use PWM, which varies the speed of the motors of the devices, so they consume only as much energy as they need [10]. 4.2.
(5) (6) where – the number characterizing the slope of the graph (the larger the , the greater the slope), – inflection point of the function ( ( ) = 0,5).
Furthermore, we will use the bell-shaped membership functions to model the mean values of fuzzy terms: ( ) = 1 + exp[− ( − )]
, ( ) = 1 + | − |2 , 1 1 where c – central (modal) value at which ( ) = 1, ≥ 0 – number characterizing the slope of the graph (similar to sigmoidal membership functions), > 0 – the distance from the center c to the inflection points of the function, where at = 0,5 the ( ± ) = 0,5 is fulfilled.
= { 2 = { ( ) = 1 1 + 0,03 −24 + 0,5 0 400
LOW
HEAVY 700 1000 1300 1600 2( ) Linguistic variable "relative humidity" and its membership function graph (Figure 7): = { through a PWM signal, consisting of 18 fuzzy rules. It is also important to note that the maximum number of rules of a fuzzy system can be found according to (7).
= ∏ where variables.
– returns the number of fuzzy terms of a linguistic variable , – number of input 0,5
LOW 25 50 100 (%)
75 2 ( ) EXCELLENT EXCELLENT
GOOD GOOD HEAVY
HEAVY EXCELLENT EXCELLENT
GOOD GOOD HEAVY
HEAVY EXCELLENT EXCELLENT
GOOD GOOD HEAVY HEAVY signal in the range from 0 to 255. Among the advantages of this system, it is worth to mention the high operating speed (no defuzzification), as well as the simplicity of training the system, which essentially boils down to finding the weights of the input parameters in the consequent of fuzzy rules and which, if necessary, can also be produced by a subject matter expert.
A fuzzy PWM controller based on the Takagi-Sugeno fuzzy inference algorithm is proposed. In this paper, a method of fuzzy control of the output PWM signal based on the Takagi-Sugeno fuzzy inference algorithm is presented. An IoT system of an intelligent fan has been developed based on the approach of converting fuzzy input parameters read from sensors into a PWM signal to control the fan screw rotation speed. A mathematical model of an intelligent fan based on fuzzy control has been developed, as well as its hardware architecture. In this approach, the model operates with three input parameters, namely, temperature, relative humidity and carbon dioxide concentrations (CO2) under different operating conditions vis-à-vis the PWM signal output. Experimental studies have been carried out that demonstrate the characteristics of the proposed methods for intelligent control of
[1] Oracle - What Is the Internet of Things (IoT)? 2021. URL: https://www.oracle.com/internet-ofthings/what-is-iot/ [2] L. Zadeh, The concept of a linguistic variable and its application to approximate reasoning.
[3] Air Conditioning. URL: https://www.energy.gov/energysaver/home-cooling-systems/airconditioning [4] Air conditioning in nearly 100 million U.S. homes, 2011. URL: https://www.eia.gov/consumption/residential/reports/2009/air-conditioning.php [5] D. Westphalen and S. Koszalinski. Energy Consumption Characteristics of Commercial Building
[6] N. Bhagwat, S.N. Teli, P. Gunaki and V.S. Majali. Review paper on energy efficiency technologies for heating, ventilation and air conditioning (HVAC). International Journal of Scientific & Engineering Research 6 (12) (2015) 106–116. [7] IoT and Machine Learning to Reduce Energy Use in Cooling Systems, 2018. URL: https://www.ibm.com/blogs/research/2018/07/reduce-energy-cooling/ [8] Fundamentals of HVAC Controls. URL: http://people.eecs.berkeley.edu/~culler/cs294f09/m197content.pdf [9] Takagi T. and Sugeno M. “Fuzzy identification of systems and its application to modeling and control.” IEEE Trans. of Systems, Man and Cybernetics 15(1) (1985): 116–132. [10] N. Brown. Introduction To PWM: How Pulse Width Modulation Works, 2021. URL: https://www.kompulsa.com/introduction-pwm-pulse-width-modulation-works/ [11] F. M. McNeill and E. Thro (1994) Fuzzy logic. A practical approach. - Academic press. - 312 p. [12] S. Yershov and R. Ponomarenko, Software architecture of hierarchical fuzzy inference.
International Conference of Programming – UkrPROG'2018. In: CEUR Workshop Proceedings, CEUR-WS.org, 2–3, 2018, pp. 99–108. [13] S. Yershov and R. Ponomarenko, Parallel Fuzzy Inference Method for Higher Order Takagi–
Sugeno Systems. Cybernetics and Systems Analysis 54(6) (2018) 170–180. [14] R. Ponomarenko. Systems for checking and testing the quality of knowledge based on fuzzy inference. CEUR Workshop Proceedings (CEUR-WS.org) “VII International conference “Information Technology and Interactions” (IT&I-2020) December 02-04”, 2021, 65-74, ISSN 1613-0073. URL: http://ceur-ws.org/Vol-2845/Paper_7.pdf [15] P. Lizunov, A. Biloshchytskyi, A. Kuchansky, Y. Andrashko and S. Biloshchytska. Improvement of the method for scientific publications clustering based on N-gram analysis and fuzzy method for selecting research partners. Eastern-European Journal of Enterprise Technologies. 4(4-100) (2019) 6–14 doi: 10.15587/1729-4061.2019.175139 [16] N. Baccar and R. Bouallegue. (2016) Interval type 2 fuzzy localization for wireless sensor networks. EURASIP J. Adv. Signal Process. 2016, 42 doi: 10.1186/s13634-016-0340-4 [17] A. Patel, A.R. Patel, D.R. Vyas and K.M. Patel. Use of PWM Techniques for Power Quality
[18] A.-H. Attia, S.F. Rezeka and A.M. Saleh. Fuzzy logic control of air-conditioning system in residential buildings. Alexandria Engineering Journal 54 (3) (2015) 395–403. doi: 10.1016/j.aej.2015.03.023 [19] S.K. Dash, G. Mohanty and A. Mohanty. Intelligent Air Conditioning System using Fuzzy Logic.
[20] M. Fayaz and D.-H. Kim. An efficient actuator control mechanism using fuzzy logic on embedded system. International Journal of Control and Automation 11(6) (2018) 35-44 doi: 10.14257/ijca.2018.11.6.04 [21] A. Sancho-Royo and J.L. Verdegay. Methods for the Construction of Membership Functions.
International Journal of Intelligent Systems 14 (12) (1999) 1213–1230 doi:10.1002/(SICI)1098111X(199912)14:123.0.CO;2-5