Advance Trends in Soft Computing

This paper presents the development of the neuro-fuzzy mathematical model of the ecopyrogenesis (EPG) complex multiloop circulatory system (MCS). The synthesis procedure of the neuro-fuzzy model, including its adaptive-network-based fuzzy inference system for temperature calculating (ANFISTC) training particularities with input variables membership functions of different types is presented. The analysis of computer simulation results in the form of static and dynamic characteristics graphs of the MCS as a temperature control object confirms the high adequacy of the developed model to the real processes. The developed neuro-fuzzy mathematical model gives the opportunity to investigate the behavior of the temperature control object in steady and transient modes, in particular, to synthesize and adjust the temperature controller of the MCS temperature automatic control system (ACS).

The estimation of fuzzy validity (

f

-validity) of complex fuzzy objects (

f

-objects) by using fuzzy geometry (f-geometry) may be a useful tool in revealing unknown links or patterns e.g. finger prints, shoe print, face sketch of a criminal etc. at crime site. The Extended Fuzzy Logic (FLe) is a combination of Fuzzy Logic (FL) and Unprecisiated Fuzzy Logic (FLu). Whenever a precise solution of any problem is either impossible or bit costlier, then we opt for the concept of FLe. The f-geometry is an example of Unprecisiated Fuzzy Logic (FLu). The

f

-geometry has different

f

-objects like

f

-point,

f

-line,

f

-circle,

f

-triangle, etc. The aggregation models can be used for aggregating the component of

f

-objects. The Minimizing Distance from extreme Point (MDP), which is a nonlinear ordered weighted averaging (OWA) objective model, is used to estimate

f

-validity of fuzzy objects. The results generated by the MDP model are found closer to degree of OR-ness. The objective of this paper is to lay the foundation and encourage further discussion on the development of‘ methods for defining as well as estimating

f

-validity of some more complex

f

-objects for forensic investigation services.

Measurements are never absolutely accurate; so, it is important to estimate how the measurement uncertainty affects the result of data processing. Traditionally, this problem is solved under the assumption that the probability distributions of measurement errors are normal – or at least are concentrated, with high certainty, on a reasonably small interval. In practice, the distribution of measurement errors is sometimes heavy-tailed, when very large values have a reasonable probability. In this paper, we analyze the corresponding problem of estimating the tail of the result of data processing in such situations.

In various works, L.A. Zadeh has introduced fuzzy quantifiers, fuzzy usuality modifiers, and fuzzy likelihood modifiers. This paper provides these notions with a unified semantics and uses this to define a formal logic capable of expressing and validating arguments such as

Most

birds can fly; Tweety is a bird; therefore, it is

likely

that Tweety can fly’. In effect, these are classical Aristotelean syllogisms that have been ‘qualified’ through the use of fuzzy quantifiers. It is briefly outlined how these, together with some likelihood combination rules, can be used to address some well-known problems in the theory of nonmonotonic reasoning.

Full exploration of databases requires advanced querying facilities. This is especially the case if user preferences related to expected results are complex. Traditional query languages like SQL and OQL only have limited facilities for expressing query criteria that are composed of simple criteria. So, while searching for information, users often have to translate their complex requirements (which are typical for human reasoning) into simpler queries, which in many cases can only partly reflect what the user is actually looking for. As a potential solution, we recently proposed a query language extension that is based on soft computing techniques and supports the use of so-called criterion trees. In this paper, we further extend criterion trees so that they can contain both mandatory and optional query conditions. More specifically, we study optional query conditions from a bipolar point of view and propose and illustrate a framework for handling them in query processing.

In climate science, knowledge about the system mostly relies on measured time series. A common problem of highest interest is the analysis of high-dimensional time series having different phases. Clustering in a multi-dimensional non-stationary time series is challenging since the problem is ill- posed. In this paper, the Finite Element Method of non-stationary clustering is applied to find regimes and the long-term trends in a temperature time series. One of the important attributes of this method is that it does not depend on any statistical assumption and therefore local stationarity of time series is not necessary. Results represent low-frequency variability of temperature and spatiotemporal pattern of climate change in an area despite higher frequency harmonics in time series.

Fuzzy community detection in social networks has caught researchers’ attention because, in most real world networks, the vertices (i.e., people) do not belong to only one community. Our recent work on generalized modularity motivated us to introduce a generalized fuzzy t-norm formulation of fuzzy modularity. We investigated four fuzzy t-norm operators, Product, Drastic, Lukasiewicz and Minimum, and the generalized Yager operator, with five well-known social network data sets. The experiments show that the Yager operator with a proper parameter value performs better than the product operator in revealing community structure: (1) the Yager operator can provide a more certain visualization of the number of communities for simple networks; (2) it can find a relatively small-sized community in a flat network; (3) it can detect communities in networks with hierarchical structures; and (4) it can uncover several reasonable covers in a complicated network. These findings lead us to believe that the Yager operator can play a big role in fuzzy community detection. Our future work is to build a theoretical relation between the Yager operator and different types of networks.

In this paper we present a decision support system that uses soft computing models for evaluation, selection and pricing of homes. The system (called LSPhome) is based on the Logic Scoring of Preference (LSP) evaluation method and implemented in the context of online real estate. The goal of this system is to use weighted compensative logic models that can precisely express user needs, and help both buyers and sellers of homes. The design of such a system creates specific logic and computational challenges. Soft computing logic problems include the use of verbalized importance scales for derivation of andness, penalty-controlled missingness-tolerant logic aggregation, detailed and verbalized presentation of evaluation results, and development of optimum pricing models. Computational problems include fast and parallel collection of heterogeneous information from the Internet, and development of user interface for fast and simple creation of customized soft computing decision criteria by nonprofessional decision makers.

Both the fuzzy measure and integral have been widely studied for multi-source information fusion. A number of researchers have proposed optimization techniques to learn a fuzzy measure from training data. In part, this task is difficult as the fuzzy measure can have a large number of free parameters (2

N

 − 2 for

N

sources) and it has many (monotonicity) constraints. In this paper, a new genetic algorithm approach to constraint preserving optimization of the fuzzy measure is present for the task of learning and fusing different ontology matching results. Preliminary results are presented to show the stability of the leaning algorithm and its effectiveness compared to existing approaches.

One of the important properties of SOM is its topology preservation of the input data. The topographic error is one of the techniques proposed to measure how well the continuity of the map is preserved. However, this topographic error is only applicable to the crisp SOM algorithms and cannot be adapted to the fuzzy SOM (FSOM) since FSOM does not assign a unique winning neuron to the input patterns. In this paper, we propose a new technique to measure the topology preservation of the FSOM algorithms. The new measure relies on the distribution of the membership values on the map. A low topographic error is achieved when neighboring neurons share similar or same membership values to a given input pattern.

The paper presents the design of type-2 fuzzy controllers using the

fuzzy Lyapunov synthesis approach

in order to systematically generate the rule base. To construct the rule base, the error signal and the derivative of the error signal are considered. It also presents the performance analysis to determine the value of the separation interval

ξ

between the upper and lower membership functions of the type-2 fuzzy set used. The controllers are implemented via simulation to solve trajectory tracking problem for angular position and angular velocity of a servo trainer equipment. Simulation results are successful for both cases and shown better performance than those of classical controllers.

In this work it is presented the design of a decentralized adaptive fuzzy control. In this scheme it is suppose a system with unknown parameters and on-line fuzzy identifier, which uses an adaptive law to adjust the unknown parameters in order to build a model of an assumed unknown nonlinear system. The applicability of the proposed approach is illustrated via simulations by trajectory tracking control of a two degrees-of-freedom robot manipulator.

The use of proposed technology is oriented for

persons

, who want

to interact

with systems

to make relevant decisions

in real-time,

fuzzy conditions

, heterogeneous

subject areas

and

multi-lingual communication

, where the

situations are unknown in advance

,

fuzzy structured

and

not clearly regulated

. The

essence

of the technology consists in the

situational control

of fuzzy

data

,

information

and

knowledge

, extracted from texts in different

natural languages

, dissimilar

subject areas

for

situational fuzzy control

of the object in

Intelligent

real-time system. The technology is formalized using

Fuzzy Logic

,

Situational Control

theories and is defined by methods of

knowledge representation

,

situational data control

,

fuzzy logic inference

,

knowledge modeling

,

generalization

and

explanation knowledge

,

dialogue control

,

machine translation

and others.

Knowledge integration is based upon gathering and aggregating all available data, information, and knowledge from theory, experience, computation and similar applications. Such a ”waste nothing” approach becomes important when the underlying theory is difficult to model, when observational data are sparse or difficult to measure, or when uncertainties are large. An inference approach is prescribed, providing common ground for many kinds of uncertainties arising from the sources of data, information and knowledge. These sources are integrated using a modified Saaty’s Analytic Hierarchy Process (AHP). A fusion physics application illustrates how to manage the uncertainties in the inference-based integration approach. Zadeh membership functions and possibility distributions contribute to this management.

Natural Language Processing (NLP) includes Tasks such as Information Extraction (IE), text summarization, and question and answering, all of which require identifying all the information about an entity exists in the discourse. Therefore a system capable of studying Co-reference Resolution (CR) will contribute to the successful completion of these Tasks. In this paper we are going to study process of Co-reference Resolution and represent a system capable of identifying Co-reference mentions for first the time in Farsi corpora. So we should consider three main steps of Farsi Corpus with Co-reference annotation, system of Mention Recognition and its domain, and the algorithm of predicting Co-reference Mentions as the basis of our study. Therefore, in first step, we prepare a Corpus with suitable labels, and this Corpus as first Farsi corpus having Mention and Co-reference labels can be the basis of many researches related to mention Detection (MD) and CR. Also using such corpus and studying rules and priorities among the mentions, we present a system that identifies the mentions and negative and positive examples. Then by using learning algorithm such as SVM, Neural Network and Decision Tree on extracted samples we have evaluated models for predicting Co-reference mentions in Farsi Language. Finally, we conclude that the performance of neural network is better than other learners.

This paper presents the real-time application of a discrete-time inverse optimal control to a three-phase linear induction motor (LIM) in order to achieve trajectory tracking of a position reference. A recurrent high-order neural network (RHONN) is employed on-line to determine the model of the motor. The equipment and software employed are described as well as real-time trajectory tracking results.

We introduce a new structure for fuzzy cognitive maps (FCM) where the traditional fan-in structure involving an inner product followed by a squashing function to describe the causal influences of antecedent nodes to a particular consequent node is replaced with a weighted mean type operator. In this paper, we employ the weighted power mean (WPM). Through appropriate selection of the weights and exponents in the WPM operators, we can both account for the relative importance of different antecedent nodes in the dynamics of a particular node, as well as take a perspective ranging continuously from the most pessimistic (minimum) to the most optimistic (maximum) on the normalized aggregation of antecedents for each node. We consider this FCM structure to be more intuitive than the traditional one, as the nonlinearity involved in the WPM is more scrutable with regard to the aggregation of its inputs. We provide examples of this new FCM structure to illustrate its behavior, including convergence.

We present a sport skill data analysis with time series image data retrieved from motion pictures, focused on table tennis. We do not use body nor skeleton model, but use only hi-speed motion pictures, from which time series data are obtained and analyzed using data mining methods such as C4.5 and so on. We identify internal models for technical skills as evaluation skillfulness for forehand stroke of table tennis, and discuss mono and meta-functional skills for improving skills.

Graph search based path planning is popular in mobile robot applications and video game programming. Previously, we developed the A-r-Star pathfinder, a suboptimal variant of the A-Star pathfinder with performance that scales linearly with increasing the resolution (size) and hence sparseness of the grid map of a given continuous world. This paper presents the study of the direct acyclic graph (tree structure) formed by the A-r-Star and outlines steps to developing an incremental version of the A-r-Star. The incremental version of A-r-Star is able to replan faster using information from previous searches to speed up subsequent searches.

The analysis of existing methods and approaches for solving transport logistics problems was performed in this paper, particularly, for optimal choice of transport company. In the working process the complex of decision making criteria was formed and the hierarchical structure of decision support system (DSS) for corresponding tasks was made. Thereby the list of different-type methods (classical and fuzzy) for synthesis of developed DSS was defined. A comparative analysis of the application of fuzzy analytic hierarchy process and the method based on fuzzy inference was held for synthesis DSS for the optimal choice of transport company. The final results prove the effectiveness and reasonability of using fuzzy modeling in problems of transport logistics.

Big data is a collection of very large and complex data that is difficult to load into the computer memory. The major challenges include searching, categorization and analysis of big data. In this paper, a fuzzy based supervised classifier is proposed to handle the searching, storage and categorization of big data. In this classifier, we proposed a Random Sampling Iterative Optimization Fuzzy c-Means (RSIO-FCM) clustering algorithm which partitions the big data into various subsets. These subsets adequately cover all the instances (object space) of big data. Then, clustering is performed on these subsets by feeding forward the centers of clustered subset to group remaining subsets. Further, the designed classifier based on Bayesian theory is used to assign the labels to these clusters and also used to predict labels of unknown instances. Thus, the proposed approach results in effective clusters formation which also eliminates the problem of overlapping cluster centers faced by algorithm discussed in [1] named as Simple Random Sampling plus Extension FCM (rseFCM). The effectiveness of proposed clustering algorithm over rseFCM clustering is evaluated on two very large benchmark datasets in terms of fuzzification parameter

m

, objective function, computational time and accuracy. Experimental results demonstrate that, the RSIO-FCM algorithm generates more appropriate cluster centers location due to which it achieves better classification accuracy as compared to the rseFCM algorithm. Thus, it observed that, cluster centers location will have significant impact over classification results.

Talent selection in cricket is a task which is usually carried out by coaches and senior players. The method relies on instincts or natural abilities of the selectors for talent assessment and selection. However, it suffers with subjectivity, personal biasness and external influences. In country such as India where more than 1-million players play cricket daily, talent selection problem becomes significant. In this paper, we propose a model which can rank players in order of their talent. The model can potentially help reduce the implicit problems of manual talent selection system. The model assesses the cricketing talent of individual players based on the quantitative outcome of the identified parametric tests for assessing players’ physical/motor, anthropometric and cognitive skills and capabilities with respect to cricket. The Ordered weighted averaging aggregation (OWA) operator with Relative Fuzzy Linguistic Quantifier (RFLQ) is used to measure the weights and aggregate players’ talent values. The model is applied to the Jamia Millia Islamia’s (JMI) University Cricket team and results have been summarized.

With the spreading of intelligent machines, man-machine communication has become an important research area. Today, intelligent robots co-operating with humans usually have to be able to store, retrieve, and update information about their environment, interpret and execute commands, offer existing and gain/learn new services. In these processes, the efficient knowledge representation and storage are of key importance. In this paper, a new graph based modular knowledge storage and representation form is presented which is able to handle inaccurate and ambiguous information, to store, retrieve, modify, and extend theoretical and practical knowledge, to interpret commands, and to learn new cognitions.

As people move through uncertain environments, they are often presented with multiple route choices. Deciding which route to take requires an understanding of the environmental features and how they affect the person’s interpreted cost of each route. These quantities can be appropriately modeled as fuzzy numbers to capture the inherent uncertainty in human knowledge. We present an approach to guide a person’s decision-making process through an environment modeled as a fuzzy weighted graph, using an

α

-level OWA operator to implement the principle of bounded rationality. A cost value is computed for each possible route choice, which can then be used to rank the set of routes and make a decision.

Fuzzy measures

(FMs) encode the

worth

(or importance) of different subsets of information sources in the

fuzzy integral

(FI). It is well-known that the

Choquet FI

(CFI) often reduces to an elementary aggregation operator for different selections of the FM. However, FMs are often learned from training data or they are derived from the densities (worth of just the singletons). In these situations an important question arises; what is the resultant CFI really doing? Is it aggregating data relative to an additive measure, a possibility measure, or something more complex and unique? Herein, we introduce new indices (distance formulas) and fuzzy sets that capture the degree to which the CFI is behaving like a set of known aggregation operators. This has practical application in terms of gaining a deeper understanding into a given problem, guiding new learning methods and evaluating the CFI’s benefit.

A methodology for modeling the electrocoagulation of wastewater from the food industres, with high organic loads is proposed. The approach used is a nonlinear model based on Artificial Neural Networks (ANN), which is able to understand the interaction between the variables that define the process, to complement the traditional design of experiments. Where the interaction of variables determines in many cases, a large number of experiments to perform, that involve stages such as planning, organization and execution of experimental activities, also characterization and analysis of wastewater in order to remove chemical oxygen demand (COD) and total dissolved solids (TSS). From this approach it will be possible to find appropriate conditions for these parameters in order to enhance the contaminant removal process with specific routes (experimental conditions).

Modern information technologies enable us to analyze great amounts of primary non-aggregated data. Publishing them increases threats of disclosing sensitive information. To protect information about a single person, one needs to provide individual data anonymity. Providing group data anonymity presupposes protecting intrinsic data features, properties, and distributions. Methods for providing group anonymity need to protect the underlying data distribution, and also to ensure sufficient data utility after their transformation. In our opinion, the latter task is a problem which can be solved using only exhaustive search, therefore heuristic procedures need to be developed to find suboptimal solutions.

Evolutionary algorithms are heuristic guided random search techniques mimicking biological evolution by natural selection. They are inherently stochastic, which turns out to be a downside when converging to an optimum. Memetic algorithms are a combination of evolutionary algorithms and local search procedures. Applying local search increases convergence and enhances algorithm performance by incorporating problem specific knowledge.

In the paper, we introduce a memetic algorithm for providing group anonymity. We illustrate its application by solving a real data based problem of protecting military personnel regional distribution.

In the present paper we construct a higher order Takagi-Sugeno fuzzy system that approximates a Mamdani fuzzy system, with arbitrary accuracy. The goal of this construction is to reduce the computational complexity of a fuzzy systems considered, also to replace a nonlinear operator by an approximate linear operator. The proposed methodology is fully constructive, so it does not require training of the Takagi-Sugeno fuzzy system. The construction combines Takagi-Sugeno systems with the classical Lagrange interpolation.

The application of soft computing in image/signal enhancement and comparing it with traditional methods will be discussed in this paper. This study presents two optimization methods for

α

-rooting image enhancement, which is a transform based method. The first method is a derivative-based optimization and the second one is Genetic Algorithm optimization. The parameter will be driven through optimization of measure of enhancement function (EME). The results from, the simulations show both methods are reliable; however, the first method has more computing cost.

This work presents an alternative approach (Genetic Algorithms approach) to traditional treatment of Recommender Systems (RSs). The work examines genetic algorithms possibilities to offer adaptive characteristics to these systems trough learning. The main goal, in addition to give a general view about RSs capabilities and possibilities, it is to provide a new example mechanism for to extend RSs learning capabilities (from user’s personal characteristics), with the purpose of improve the effectiveness at time of to find recommendations and appropriate suggestions for particular individuals.

Fuzzy sets have been originally introduced as

generalizations

of crisp sets, and this is how they are usually considered. From the mathematical viewpoint, the problem with this approach is that most notions allow many different generalizations, so every time we try to generalize some notions to fuzzy sets, we have numerous alternatives. In this paper, we show that fuzzy sets can be alternatively viewed as

limits

of crisp sets. As a result, for some notions, we can come up with a unique generalization – as the limit of the results of applying this notion to the corresponding crisp sets.

To properly process data, we need to know the accuracy of different data points, i.e., accuracy of different measurement results and expert estimates. Often, this accuracy is not given. For such situations, we describe how this accuracy can be estimated based on the available data.

This paper is a contribution to the Structural Health Monitoring problem, solved by using case based reasoning and Self Organizing Maps. The expert system described in this paper is able to detect, locate and quantify stiffness percentage changes in a mechanical engineering structure. In order to overcome issues relating large number of parameters involved in the training stage it was applyed differential evolutive algorithms. Proper indexes to evaluate the training quality were proposed in order to increase diagnosis reliability. The algorithms were tested using the UBC ASCE Benchmark. The numerical implementation shows decreasing in the identification errors with respect to those obtained by selecting manually network training parameters.

In many practical situations, it is necessary to cluster given situations, i.e., to divide them into groups so that situations within each group are similar to each other. This is how we humans usually make decisions: instead of taking into account all the tiny details of a situation, we classify the situation into one of the few groups, and then make a decision depending on the group containing a given situation. When we have many situations, we can describe the probability density of different situations. In terms of this density, clusters are connected sets with higher density separated by sets of smaller density. It is therefore reasonable to define clusters as connected components of the set of all the situations in which the density exceeds a certain threshold

t

. This idea indeed leads to reasonable clustering. It turns out that the resulting clustering works best if we use a Gaussian function for smoothing when estimating the density, and we select a threshold in a certain way. In this paper, we provide a theoretical explanation for this empirical optimality. We also show how the above clustering algorithm can be modified so that it takes into account that we are not absolutely sure whether each observed situation is of the type in which we are interested, and takes into account that some situations “almost” belong to a cluster.

In many real-life situations, e.g., in medicine, it is necessary to process data while preserving the patients’ confidentiality. One of the most efficient methods of preserving privacy is to replace the exact values with intervals that contain these values. For example, instead of an exact age, a privacy-protected database only contains the information that the age is, e.g., between 10 and 20, or between 20 and 30, etc. Based on this data, it is important to compute correlation and covariance between different quantities. For privacy-protected data, different values from the intervals lead, in general, to different estimates for the desired statistical characteristic. Our objective is then to compute the range of possible values of these estimates.

Algorithms for effectively computing such ranges have been developed for situations when intervals come from the original surveys, e.g., when a person fills in whether his or her age is between 10 or 20, between 20 and 30, etc. These intervals, however, do not always lead to an optimal privacy protection; it turns out that more complex, computer-generated “intervalization” can lead to better privacy under the same accuracy, or, alternatively, to more accurate estimates of statistical characteristics under the same privacy constraints. In this paper, we extend the existing efficient algorithms for computing covariance and correlation based on privacy-protected data to this more general case of interval data.

Process interruptions are carried out either automatically by monitoring and control systems that react to deviations from standards or by operators reacting to anomalies or incidents. Process interruptions in (very) large production systems are difficult to trace and to deal with; an extended stop is also very costly and solutions are sought to find an effective support technology to minimize the number of involuntary process interruptions. Feature selection is intended to reduce the complexity of handling the interactions of numerous factors in large process systems and to help find the best ways to handle process interruptions. We show that feature selection can be carried out with fuzzy entropy and interval-valued fuzzy sets.

By their widespread application, engineering virtual spaces together with computation methods play key role in product information management in order to assist product definition and the relevant decisions. Research in order to include hard and soft computing methods in well organized product model of increasing intelligence is a main objective of product model development efforts. This paper is a contribution to these efforts by a possible method for including a higher level knowledge based modeling as extension to the currently applied product models. After an introduction of relevant characteristics of engineering problem solving, role of soft computing is discussed. Following this, extending of knowledge integration in product model by the proposed method is explained and new features for the extended product model are introduced. Finally, the proposed new context structure is outlined and its implementation in product lifecycle management (PLM) systems is conceptualized.

This contribution presents an infinite-horizon optimal tracking controller for nonlinear systems based on the state-dependent Riccati equation approach. The synthesized control law comes from solving the Hamilton-Jacobi-Bellman equation for state-dependent coefficient factorized (SDCF) nonlinear systems. The proposed controller minimizes a quadratic performance index, whose entries are determined by the particle swarm optimization (PSO) algorithm in order to improve the performance of the control system by fulfilling with design specifications such as bound of the control input expenditure, steady-state tracking error and rise time. The effectiveness of the proposed PSO optimal tracking controller is applied via simulation to the Van der Pol Oscillator.

In clinical practice, making diagnostically crisp decisions is critical to successful treatment outcomes. However, there is no agreement on the operational methodology that is best suited to convert imprecise symptomatic information into crisp clinical treatment decision making. In this paper, a new computational decision making tool, Delphi-Neural Decision Making Processor (D-NDMP), is introduced as a preliminary study to apply to clinical practices for more successful and efficient operational decisions. A case study in a dental clinical decision involving a deep decay tooth is presented as an example to perform D-NDMP. The results yield a more reliable and confident opinion on the practical application of treatment decision in uncertain cases in a clinical decision making process.

A widespread and growing use of information systems, notably databases, calls for formalisms to specify user preferences richer than classical query languages. The concept of bipolarity of preferences is recently considered crucial in this respect. Its essence consists in considering separately positive and negative evaluations provided by the user which are not necessarily a complement of each other. In our previous work we have proposed an approach based on a specific interpretation of the positive and negative evaluations and their combination using a non-standard logical connective referred to as “and possibly”. Here we propose a novel extension of that approach. We present the concepts, possible interpretations and some analyses.

In this paper, we present control systems for an unmanned aerial vehicle (UAV) which provides aerial support for an unmanned ground vehicle (UGV). The UGV acts as a mobile launching pad for the UAV. The UAV provides additional environmental image feedback to the UGV. Our UAV of choice is a Parrot ArDrone 2.0 quadcopter, a small four rotored aerial vehicle, picked for its agile flight and video feedback capabilities. This paper presents design and simulation of fuzzy logic controllers for performing landing, hovering, and altitude control. Image processing and Mamdani-type inference are used for converting sensor input into control signals used to control the UAV.

Many qualitative group decisions in professional fields such as law, engineering, economics, psychology, and medicine that appear to be crisp and certain are in reality shrouded in fuzziness as a result of uncertain environments and the nature of human cognition within which the group decisions are made. In this paper we introduce an innovative approach to group decision making in uncertain situations by using fuzzy theory and a mean-variance neural approach. The key idea of this proposed approach is to defuzzify the fuzziness of the evaluation values from a group, compute the excluded-mean of individual evaluations and weight it by applying a variance influence function (VIF); this process of weighting the excluded-mean by VIF provides an improved result in the group decision making.

We present an iterative preprocessing approach for training a support vector machine for a large dataset, based on balancing the center of mass of input data within a variable margin about the hyperplane. At each iteration, the input data is projected on the hyperplane, and the imbalance of the center of mass for different classes within a variable margin is used to update the direction of the hyperplane within the feature space. The approach provides an estimate for the margin and the regularization constant. In the case of fuzzy membership of the data, the membership function of the input data is used to determine center of mass and to count data points which violate the margin. An extension of this approach to non-linear SVM is suggested based on dimension estimation of the feature space represented via a set of orthonormal feature vectors.