Fuzzy Inference System Structure Research Articles

Implementation of Sliding Surface Moving Anfis Based Sliding Mode Control to Rotary Inverted Pendulum

This study covers the control of the pendulum angle by taking into account the dynamic equations and motor dynamics of the rotary inverted pendulum system, with the help of state variables in the Matlab program, by using the sliding mode control method with sliding surface moving and the adaptive neural fuzzy inference system together. The sliding mode control method with a changing sliding surface is a part of the control structure. The slope of the sliding surface was calculated using the adaptive neural fuzzy inference technique. The optimum values of the coefficients in the adaptive neural-fuzzy inference system structure have been calculated by genetic algorithm. The finding of the coefficients, the sum of the squares of the errors chosen as the objective function. The input of the adaptive neural fuzzy inference system structure consists of the error of the pendulum and the derivative of the error of the pendulum. The gradient of the sliding surface of the sliding mode control structure is the output of the adaptive neural fuzzy inference system structure. According to the findings, the pendulum angle achieved the appropriate reference value after 1.5 seconds, with an error of around zero. It obtained that the engine torque value reaches up to 50 Nm. From here, it is seen that the motor torque values used in practical applications and the motor torque values as a result of this study overlap.

Neurofuzzy modelling on the influence of Pt–Sn catalyst properties in direct ethanol fuel cells performance: Fuzzy inference system generation and cell power density optimization

Pt-based materials are commonly employed as catalysts in the electrooxidation of ethanol in direct ethanol fuel cells (DEFC). However, due to the complex nature of electrooxidation, the effect of catalyst properties on fuel cell power generation is not easily described by mechanistic models. Fuzzy systems can be used instead precisely for their ability to provide a more direct and qualitative/semiquantitative cause-effect relationship. In this study, a neurofuzzy model was developed to relate the effect of 5 input variables (4 catalyst properties, namely, crystal size, surface area, presence of PtSn phase and Pt L3-edge whiteline integrated intensity, and the cell potential) on cell current density. The fuzzy inference system (FIS) structure was constructed in MATLAB with ANFIS (Adaptive network-based FIS) based on experimental data used for training and validation. A 4-variable FIS (not including integrated intensity) was initially created and used as a basis for developing the 5-variable FIS, thus addressing the issue of fewer experimental data points available for integrated intensity. Both FIS structures exhibited very good fits to experimental data. The addition of integrated intensity as an input variable did not affect the fitted fuzzy model parameters for the other input variables and, therefore, a broader and more relevant model that included an electronic property of the catalyst was created. Response surface analyses, corroborated by particle swarm optimization, indicated that decreasing crystal size has the greatest effect in maximizing power density. Medium potentials and medium integrated intensity are also favorable. In addition, presence of the PtSn phase in moderate amounts is not unfavorable. With these values it was possible to predict the optimization of the power density value to 24.3 mW/cm2, compared to the best experimental value of 19.6 mW/cm2.

Training fuzzy inference system-based classifiers with Krill Herd optimization

In recent years, researchers have been able to access many novel metaheuristic algorithms inspired by natural phenomena. One such bio-inspired optimization routine is Krill Herd Algorithm (KHA). In this study, a new approach for modification of membership function parameters in a fuzzy inference system (FIS) is demonstrated. Here, the main intent is to compare KHA optimization with other heuristic and metaheuristic algorithms, as a means to train FIS structures. The proposed FIS training method has been designed to serve as a fuzzy classifier. Hence, benchmark data sets extracted from the University of California, Irvine (UCI) Machine Learning Repository were applied, while Classification Errors and Sum of Squared Errors were used as measures for evaluation criteria. The obtained results led to the conclusion that the utilization of KHA provides promising performance, especially in the case of imbalanced data—whether in terms of the classification measures or the time required for an adequate FIS training.

Adaptive Neuro-Fuzzy Inference System for Prediction of Surgery Time for Ischemic Stroke Patients

With the advent of machine learning techniques, creation and utilization of prediction models for different medical procedures including prediction of diagnosis, treatment and recovery of different medical conditions has become the norm. Recent studies focus on the automation of infarction volume growth rate prediction by the utilization of machine learning techniques. These techniques when effectively applied, could significantly help in reducing the time needed to attend to stroke patients. We propose, in this proposal, a Fuzzy Inference System that can determine when a stroke patient should undergo Decompressive Hemicraniectomy. The second infarction volume growth rate and the decision whether a patient needs to undergo this procedure, both predicted outputs of two trained models, act as inputs to this system. While the initial prediction model, that which predicts the second infarction volume growth rate is adopted from an earlier model, we propose the later model in this paper. Three Machine Learning techniques - Support Vector Machine, Artificial Neural Network and Adaptive Neuro Fuzzy Inference System with and without the feature reduction technique of Principle Component Analysis were modelled and evaluated, the best of which was selected to model the proposed prediction model. We also defined the structure of Fuzzy Inference System along with its rules and obtained an overall accuracy of 95.7% with a precision of 1 showing promising results from the use of fuzzy logic.

A Proposal for a Decision-Making Tool in Third-Party Logistics (3PL) Provider Selection Based on Multi-Criteria Analysis and the Fuzzy Approach

The selection of a third-party logistics (3PL) provider is an important and demanding task for many companies and organizations dealing with distribution activities. To assist in decision making, this paper proposes the implementation of fuzzy logic. To design a fuzzy inference system (FIS), the first prerequisite is to determine a set of evaluation criteria and sub-criteria and to find the relationship between them. This task was solved by an extensive review of the literature and expert opinions on implementing the Fuzzy Analytic Hierarchy Process (AHP) approach. The results obtained in the first part of the research, together with data collected from 20 3PL providers, were further used in the second part, which was related to the implementation of the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method. Finally, a decision-making tool for 3PL provider selection was designed as an FIS structure, where the inputs were the previously defined criteria and the output was a preference for 3PL selection. The fuzzy rules were generated on the basis of the collected empirical data, the preferences obtained by the TOPSIS method, and expert opinion using the Wang–Mendel method. The proposed fuzzy model is particularly suitable when input data are not crisp values but are provided descriptively through linguistic statements.

Modelling and characterisation of a magneto-rheological elastomer isolator device under impact loadings using interpolated multiple adaptive neuro fuzzy inference system structure

This paper presents the modelling and characterisation of a magneto-rheological elastomer isolator device (MREID) under impact loading using the adaptive neuro-fuzzy inference system (ANFIS) technique. The characterisation of an MREID under impact loading was performed using an impact pendulum test rig, and the data obtained from the experimental work was processed in form of force-velocity and force-displacement characteristics. In order to predict MREID behaviour in simulation analysis, multiple ANFIS models were proposed. A single ANFIS model represented a single kinetic energy produced by the impact mass used in experimental work. The experimental data was then used to train the ANFIS in predicting MREID behaviour and validating its performance. For verification, the prediction model, a parametric model (namely, the modified Bouc-Wen model) was developed, and the models were compared. The proposed interpolated multiple ANFIS model predicted the behaviour of the MREID with a high level of accuracy. The proposed model produced a better prediction than the modified Bouc-Wen model.

Modelling and characterisation of a magneto-rheological elastomer isolator device under impact loadings using interpolated multiple adaptive neuro fuzzy inference system structure

This paper presents the modelling and characterisation of a magneto-rheological elastomer isolator device (MREID) under impact loading using the adaptive neuro-fuzzy inference system (ANFIS) technique. The characterisation of an MREID under impact loading was performed using an impact pendulum test rig, and the data obtained from the experimental work was processed in form of force-velocity and force-displacement characteristics. In order to predict MREID behaviour in simulation analysis, multiple ANFIS models were proposed. A single ANFIS model represented a single kinetic energy produced by the impact mass used in experimental work. The experimental data was then used to train the ANFIS in predicting MREID behaviour and validating its performance. For verification, the prediction model, a parametric model (namely, the modified Bouc-Wen model) was developed, and the models were compared. The proposed interpolated multiple ANFIS model predicted the behaviour of the MREID with a high level of accuracy. The proposed model produced a better prediction than the modified Bouc-Wen model.

Evaluation on Partial Discharge Intensity of Electrical Equipment Based on Improved ANFIS and Ultraviolet Pulse Detection Technology

The ultraviolet pulse detection technology has superiorities in response speed and antijamming capability, so it is extensively applied to insulation detection for electrical equipment. The existing detection and analysis methods have defects, such as failure of precise modeling caused by incomplete analysis of influence factor, poor model adaptability, and complex operation. In allusion to the problems, the ultraviolet pulse detection circuit was optimized and its sensitivity was analyzed through tests. Then, a partial discharge intensity evaluation method for electrical equipment based on the improved adaptive network based fuzzy inference system (improved ANFIS) was proposed combined with the detected pulse count (P), temperature (T) and humidity (H). The initial fuzzy inference system structure was established with the subtractive clustering method (SCM) and fuzzy C-means (FCM) algorithm, and the traditional ANFIS learning algorithm was improved via Fletcher-Reeves conjugate gradient method. In this way, the model parameters were optimized continuously, and the system ability of ignoring small changes in the network was improved. Finally, the effectiveness and practicability of the method were verified through field test. The experimental results demonstrated that the improved ANFIS reduced the model error by 2% when compared with traditional ANFIS, and the model accuracy is improved. Besides, the quantitative precision of the discharge intensity is higher than that of the traditional ANFIS by the contrastive analysis of field test data, indicating that the improved ANFIS evaluate the partial discharge intensity of electrical equipment more accurately. Thus, a decision basis can be provided for the equipment protection and charged maintenance of the electric system.

Modeling and prediction of water quality parameters using a hybrid particle swarm optimization–neural fuzzy approach

This work focuses on the correlations among water quality parameters such as total hardness, total dissolved solids, electrical conductivity and total alkalinity with water parameters pH, temperature and the sum of mill equivalents of cations and anions in water reservoir in Kermanshah province (located in the middle of the western part of Iran). The data of water quality of monitoring sites were collected over year 2015. To predict and simulate water quality parameters, two data-driven models, i.e., adaptive neural fuzzy inference system and a hybrid adaptive neural fuzzy inference system structure trained by particle swarm optimization technique, were used. The main advantages of these methods are their high accuracy and very fast computational speed to predict unknown data. The results indicated that implementation of two models is highly satisfactory for predicting inorganic indicators of water quality. However, the flexibility of particle swarm optimization–adaptive neural fuzzy inference system method in modeling is better than adaptive neural fuzzy inference system approach. To prove this, the correlation coefficient, mean absolute error, root mean square error and t statistics were calculated as the error criterion. The overall (training and testing) mean relative error percentage, mean absolute error, root mean square error, correlation coefficient and t statistics obtained by the proposed particle swarm optimization–adaptive neural fuzzy inference system model are less than 3.50, 11.60, 18.90, 0.95 and 0.38%, respectively. The results provide a useful approach that uses water parameters to estimate water quality in water reservoir for water treatment and pollution management.

Determination of the ultrasound power effects on flavonoid compounds from Psidium guajava L. using ANFIS

The Adaptive Neural Fuzzy Inference System (ANFIS) model was used to predict and optimize the content of flavonoid compounds in guava leaves (Psidium Guajava L.). The extraction process was carried out by using ultrasound assisted extraction (UAE) with the variable parameters: temperature ranging from 25°C to 35°C, ultrasonic frequency (30 - 40 kHz) and extraction time (20 - 40 minutes). ANFIS learning procedure began by providing the input variable data set (temperature, frequency and time) and the output of the flavonoid compounds from the experiments that had been done. Subtractive clustering methods was used in the manufacture of FIS (fuzzy inference system) structures by varying the range of influence parameters to generate the ANFIS system. The ANFIS trainingsconducted wereaimed at minimum error value. The results showed that the best ANFIS models used a subtractive clustering method, in which the ranges of influence 0.1 were 0.70 x 10-4 for training RMSE, 8.11 for testing RMSE, 2.7 % MAPE, and 7.72 MAE. The optimum condition was obtained at a temperature of 35°C and frequency of 40 kHz, for 30 minutes. This result proves that the ANFIS model can be used to predict the content of flavonoid compounds in guava leaves.

Rainfall prediction using fuzzy inference system for preliminary micro-hydro power plant planning

East Kalimantan is a very rich area with water sources, in the form of river streams that branch to the remote areas. The conditions of natural potency like this become alternative solution for area that has not been reached by the availability of electric energy from State Electricity Company. The river water in selected location (catchment area) which is channelled to the canal, pipeline or penstock can be used to drive the waterwheel or turbine. The amount of power obtained depends on the volume/water discharge and headwater (the effective height between the reservoir and the turbine). The water discharge is strongly influenced by the amount of rainfall. Rainfall is the amount of water falling on the flat surface for a certain period measured, in units of mm3, above the horizontal surface in the absence of evaporation, run-off and infiltration. In this study, the prediction of rainfall is done in the area of East Kalimantan which has 13 watersheds which, in principle, have the potential for the construction of Micro Hydro Power Plant. Rainfall time series data is modelled by using AR (Auto Regressive) Model based on FIS (Fuzzy Inference System). The FIS structure of the training results is then used to predict the next two years rainfall.

Genetic algorithm tuned fuzzy inference system to evolve optimal groundwater extraction strategies to control saltwater intrusion in multi-layered coastal aquifers under parameter uncertainty

Excessive withdrawal of groundwater resources poses significant challenges to the management of saltwater intrusion processes in coastal aquifers. Optimization of groundwater withdrawal rates plays a vital role in sustainable management of coastal aquifers. This study proposes a genetic algorithm (GA) tuned Fuzzy Inference System (FIS) hybrid model (GA-FIS) for developing a regional scale saltwater intrusion management strategy. GA is used to tune the FIS parameters in order to obtain the optimal FIS structure. The GA-FIS models thus obtained are linked externally to the Controlled Elitist Multi-objective Genetic Algorithm (CEMGA) in order to derive optimal pumping management strategies using a linked simulation–optimization approach. The performance of the hybrid GA-FIS-CEMGA based saltwater intrusion management model is compared with that of a basic adaptive neuro fuzzy inference system (ANFIS) based management model (ANFIS-CEMGA). The parameters of the ANFIS model are tuned using hybrid algorithm. To achieve computational efficiency, the proposed optimization routine is run in a parallel processing platform. An illustrative multi-layered coastal aquifer system is used to evaluate the performances of both management models. The illustrative aquifer system considers uncertainties associated with the hydrogeological parameters e.g. hydraulic conductivity, compressibility, bulk density, and aquifer recharge. The evaluation results show that the proposed saltwater intrusion management models are able to evolve reliable optimal groundwater extraction strategies to control saltwater intrusion for the illustrative multi-layered coastal aquifer system. However, a closer look at the performance evaluation results demonstrate the superiority of the GA-FIS-CEMGA based management model over ANFIS-CEMGA based saltwater intrusion management model.

Relating rock mass properties with Lugeon value using multiple regression and nonlinear tools in an underground mine site

The Lugeon test is one of the commonly applied field methods for measuring hydraulic conductivity of a rock mass. Understanding hydraulic conductivity is especially necessary when groundwater is present, as it has a direct effect on the construction operations and stability of a structure. Discontinuity orientation, spacing and discontinuous surface quality, and the presence and type of infill play essential roles in permeability of a rock mass. Commonly used rock quality designation (RQD) and discontinuity surface condition rating of the rock mass rating system (Dc) were chosen as predictive parameters. Additionally, depth is involved as a critical predictor and it is observed so. Three variables impacting the Lugeon value are not present in the literature. The importance of each predictor variable was found to be significant while depth contributed more. Simple regression work resulted in insufficient correlation for each single parameter, but indicated they have relevance to the Lugeon value. In addition to linear and nonlinear multiple regression studies, Box-Cox transformation multiple regression was employed and predictions were found to be statistically significant. Among the multiple regression models, a nonlinear model provided the highest prediction performance. Utilization of the adaptive neuro-fuzzy inference system (ANFIS) enabled researchers to predict the Lugeon value precisely, compared to the multiple regression works. Subtractive clustering was employed in order to successfully model the parameters by using ANFIS. The clustering task resulted in a fuzzy inference system structure with three rules. A manually introduced fuzzy inference system (FIS) structure with 27 rules exhibited low performance when it was compared to the structure generated by subtractive clustering. The findings can be used in the study area since a wide range of rock types, properties and depth were taken into account in the models. Groundwater flow and permeability in jointed rock mass have a complex mechanism with variable fracturing and discontinuity properties within a small area. For prediction work, it is concluded to be beneficial to add the depth parameter to the models for further studies.

Prediction of solubility of solid compounds in supercritical CO2 using a connectionist smart technique

In this communication, an Adaptive Neuro-Fuzzy Inference System (ANFIS) was developed for estimation of solubility of several solid species in supercritical CO2. A total of 795 data points were collected from the literature and used in the process of model development and evaluation. The Genetic Algorithm (GA) was implemented to optimize the radius of influence of initial Fuzzy Inference System (FIS). The tuning parameters of FIS structure were optimized by utilizing a new method based on conjugation of hybrid and Particle Swarm Optimization (PSO) method namely CHPSO method. The accuracy of the proposed model was evaluated using graphical and statistical analysis and by comparing the results with another model reported in the literature. Results show that the developed conjugate ANFIS model is robust and precise in prediction of experimental data and is superior to other literature model.

A Comparative Study of Adaptive Neuro-Fuzzy Inference Systems in Object Detection of Complex City Scenes Using Digital Aerial Images and LiDAR Data

As adaptive neuro-fuzzy inference systems (ANFISs) have shown a high capability in solving various complicated problems, their usage have been increased so far. ANFISs are able to deal with large amounts of data with linear or nonlinear relations. In this paper, an ANFIS-based methodology was proposed in order to detect objects in complex city scenes using digital aerial images and LiDAR data. In this regard, four features were generated and normalized as ANFIS inputs including Green band, normalized difference vegetation index (NDVI), and normalized digital surface model (nDSM) using two different algorithms via morphological operations. Two types of objects were detected including buildings and trees. The proposed ANFIS used three different algorithms to build its fuzzy inference system structure including grid partition, subtractive clustering, and fuzzy c-means clustering. The results were evaluated on three different test areas provided by the fourth working group of the third commission (WG III/4) of international society of photogrammetry and remote sensing (ISPRS), over Vaihingen in Germany, known as Areas 1, 2, and 3. The achieved results were compared with each other, as well as with ISPRS WG III/4 participants’ results, by considering Completeness, Correctness, Quality, and RMS indices per-area and per-object levels. The achieved results demonstrated the capability of the proposed ANFIS in detecting buildings and trees in complex city scenes in comparison with other methods. In building detection, the proposed ANFIS-based methods achieved Completeness of 100 % in all three test areas for buildings larger than 50 m2. Also in both tree detection and building detection, the proposed ANFIS-based methods achieved Completeness and Correctness values larger than 90 %, by considering objects larger than 50 m2.

Modelling and Comparison of Compressor Performance Parameters by Using ANFIS

Developing a robust control algorithm for an aircraft engine requires an accurate nonlinear mathematical model. In formation of a nonlinear mathematical model, some components like compressor and turbine are modeled by using component maps. These maps show the connection between the compressor performance parameters. To show this connection, map data is digitized by using some techniques. In this study, we digitized a compressor map data by using ANFIS (Adaptive Neuro Fuzzy Inference System). RMSE (Root Mean Square Error) were calculated for different types of FIS (Fuzzy Inference System) structures constructed with different number of membership functions. The model was formed by using all valid data which is collected from a small turboprop engine compressor. Results demonstrate that the designed ANFIS structure can serve as an alternative model to estimate both online and offline compressor performance parameters.

A novel minimum–maximum data-clustering algorithm for vibration control of a semi-active vehicle suspension system

In this paper, a new control system for a semi-active vehicle suspension is presented by proposing a novel minimum–maximum (min–max) pure data-clustering algorithm. The min–max pure data-clustering algorithm is used to create pure data clusters which are the basic structure for building fuzzy sets and membership functions of an adaptive neuro-fuzzy inference system. Based on the data clusters created by the min–max pure data-clustering algorithm, an appropriate structure of the adaptive neuro-fuzzy inference system is established and used to identify the dynamic behaviour of a semi-active vehicle suspension system featuring a magnetorheological-fluid-based damper. In this control system, both the measured data and an inverse dynamic model of the damper are used. To calculate the desired damping force value corresponding to the road profile at a specific time, a fuzzy inference system is built on the basis of a genetic algorithm. In this work, the fitness function of the genetic algorithm is satisfactorily considered to create the optimal fuzzy inference system structure, which expresses the ride-comfort-oriented tendency. Based on the desired force value, the desired current value is obtained by the inverse dynamic model. This is the optimal current value to be applied to the magnetorheological-fluid-based damper to reduce the acceleration of the vehicle. The effectiveness of the proposed control algorithm is demonstrated by vibration control performances such as reducing the vertical acceleration of the vehicle body and increasing the road-holding ability of the vehicle tyre. In addition, a comparison between the proposed work and the previous work is undertaken in order to show the superior vibration control performance of the proposed control strategy.

Bottleneck Prediction Method Based on Improved Adaptive Network-based Fuzzy Inference System (ANFIS) in Semiconductor Manufacturing System

Semiconductor manufacturing (SM) system is one of the most complicated hybrid processes involved continuously variable dynamical systems and discrete event dynamical systems. The optimization and scheduling of semiconductor fabrication has long been a hot research direction in automation. Bottleneck is the key factor to a SM system, which seriously influences the throughput rate, cycle time, time-delivery rate, etc. Efficient prediction for the bottleneck of a SM system provides the best support for the consequent scheduling. Because categorical data (product types, releasing strategies) and numerical data (work in process, processing time, utilization rate, buffer length, etc.) have significant effect on bottleneck, an improved adaptive network-based fuzzy inference system (ANFIS) was adopted in this study to predict bottleneck since conventional neural network-based methods accommodate only numerical inputs. In this improved ANFIS, the contribution of categorical inputs to firing strength is reflected through a transformation matrix. In order to tackle high-dimensional inputs, reduce the number of fuzzy rules and obtain high prediction accuracy, a fuzzy c-means method combining binary tree linear division method was applied to identify the initial structure of fuzzy inference system. According to the experimental results, the main-bottleneck and sub-bottleneck of SM system can be predicted accurately with the proposed method.

An adaptive neuro-fuzzy approach for modeling of water-in-oil emulsion formation

Oil composition and properties including density, viscosity, asphaltene, saturate, aromatics and resin contents are responsible factors for the formation of water-in-crude-oil emulsions. These factors can be used to develop an stability index which determines states of water-in-oil emulsion in terms of either an unstable, entrained, mesostable or stable conditions. It is important to note that most of the regression models cannot capture the non-linear relationships involved in the formation of these emulsions. This study deals with the prediction of water-in-oil emulsions stability by an adaptive neuro-fuzzy inference system (ANFIS) with basic compositional factors such as density, viscosity and percentages of SARA (saturates, aromatics, resins, and asphaltenes) components. In the computational method, grid partition and subtractive clustering fuzzy inference systems were tried to generate the optimum fuzzy rule base sets. The stability estimation was conducted by applying hybrid learning algorithm and the model performance was tested by the means of distinct test data set randomly selected from the experimental domain. The ANFIS-based predictions were also compared to the conventional regression approach by means of various descriptive statistical indicators, such as root mean-square error (RMSE), index of agreement (IA), the factor of two (FA2), fractional variance (FV), proportion of systematic error (PSE), etc. With trying various types of fuzzy inference system (FIS) structures and several numbers of training epochs ranging from 1 to 100, the lowest root mean square error (RMSE = 2.0907) and the highest determination coefficient ( R 2 = 0.967) were obtained with subtractive clustering method of a first-order Sugeno type FIS. For the optimum ANFIS structure, input variables were fuzzified with four Gaussian membership functions, and the number of training epochs was computed as 21. In the computational analysis, the predictive performance of the ANFIS model was examined for the following ranges of the clustering parameters: range of influence (ROI) = 0.45–0.60, squash factor (SF) = 1.20–1.35, accept ratio (AR) = 0.40–0.55, and reject ratio (RR) = 0.10–0.20. Results indicated that ROI, SF, AR and RR were obtained to be 0.54, 1.25, 0.50 and 0.15, respectively, for the best FIS structure. It was clearly concluded that the proposed ANFIS model demonstrated a superior predictive performance on forecasting of water-in-oil emulsions stability. Findings of this study clearly indicated that the neuro-fuzzy modeling could be successfully used for predicting the stability of a specific water-in-oil mixture to provide a good discrimination between several visual stability conditions.

A new hybrid module for skin detector using fuzzy inference system structure and explicit rules

Skin detection is a popular image processing technique that has been applied in many areas such as video-surveillance, cyber-crime prosecution and face detection. It is also considered as one of the challenging problems in image processing. Despite being a well known technique to detect human appearance within image, it faces a fundamental problem when using colour as cue to detect skin. It is difficult to detect skin when the colour between the skin and the non skin within an image is similar. Therefore in this paper, a new hybrid module between explicit rules and fuzzy inference system structure, based on RGB colour space, is proposed to improve skin detection performance. Using the new hybrid module, we managed to increase the classification reliability when discriminating human skin. The new proposed skin detector depends on subtractive clustering technique, created and trained with training set of skin and non-skin pixels. The proposed system is tested on human images having upright frontal skin with any background. Our proposed system has achieved high detection rates of 87% classification and low false positives when compared with the existing methods. Key words: