Band 25, Heft 4

Aging of the population in developed countries is an important challenge to overcome. Robots have a role to play to offer solutions to old and dependent persons and also to their families and carers. We investigated low-cost high-tech robotics in order to offer affordable mobile robots that could be used in the context of ambient assisted living (AAL). We used a bottom–up approach and tried different technical solutions in order to propose an efficient prototype. Results of this experimental work are presented in the first part of this paper. From the different experiments we carried out, we felt the need to use a top–down approach that could be used to develop such robots and also any system that requires the use of mechanical, electronic, and software components. In the second part of this paper, we present the X-methodology and try to show its use in our AAL context.

Thresholding is the easiest method for image segmentation. Bi-level thresholding is used to create binary images, while multilevel thresholding determines multiple thresholds, which divide the pixels into multiple regions. Most of the bi-level thresholding methods are easily extendable to multilevel thresholding. However, the computational time will increase with the increase in the number of thresholds. To solve this problem, many researchers have used different bio-inspired metaheuristics to handle the multilevel thresholding problem. In this paper, optimal thresholds for multilevel thresholding in an image are selected by maximizing three criteria: Between-class variance, Kapur and Tsallis entropy using harmony search (HS) algorithm. The HS algorithm is an evolutionary algorithm inspired from the individual improvisation process of the musicians in order to get a better harmony in jazz music. The proposed algorithm has been tested on a standard set of images from the Berkeley Segmentation Dataset. The results are then compared with that of genetic algorithm (GA), particle swarm optimization (PSO), bacterial foraging optimization (BFO), and artificial bee colony algorithm (ABC). Results have been analyzed both qualitatively and quantitatively using the fitness value and the two popular performance measures: SSIM and FSIM indices. Experimental results have validated the efficiency of the HS algorithm and its robustness against GA, PSO, and BFO algorithms. Comparison with the well-known metaheuristic ABC algorithm indicates the equal performance for all images when the number of thresholds M is equal to two, three, four, and five. Furthermore, ABC has shown to be the most stable when the dimension of the problem is too high.

The vital signs of chronically ill patients are monitored daily. The record flags when a specific vital sign is stable or when it trends into dangerous territory. Patients also self-assess their current state of well-being, i.e. whether they are feeling worse than usual, neither unwell nor very well compared to usual, or are feeling better than usual. This paper examines whether past vital sign data can be used to forecast how well a patient is going to feel the next day. Reliable forecasting of a chronically sick patient’s likely state of health would be useful in regulating the care provided by a community nurse, scheduling care when the patient needs it most. The hypothesis is that the vital signs indicate a trend before a person feels unwell and, therefore, are lead indicators of a patient going to feel unwell. Time series and classification or regression tree methods are used to simplify the process of observing multiple measurements such as body temperature, heart rate, etc., by selecting the vital sign measures, which best forecast well-being. We use machine learning techniques to automatically find the best combination of these vital sign measurements and their rules that forecast the wellness of individual patients. The machine learning models provide rules that can be used to monitor the future wellness of a patient and regulate their care plans.

Most of the state-of-the-art speaker identification systems work on a monolingual (preferably English) scenario. Therefore, English-language autocratic countries can use the system efficiently for speaker recognition. However, there are many countries, including India, that are multilingual in nature. People in such countries have habituated to speak multiple languages. The existing speaker identification system may yield poor performance if a speaker’s train and test data are in different languages. Thus, developing a robust multilingual speaker identification system is an issue in many countries. In this work, an experimental evaluation of the modeling techniques, including self-organizing map (SOM), learning vector quantization (LVQ), and Gaussian mixture model-universal background model (GMM-UBM) classifiers for multilingual speaker identification, is presented. The monolingual and crosslingual speaker identification studies are conducted using 50 speakers of our own database. It is observed from the experimental results that the GMM-UBM classifier gives better identification performance than the SOM and LVQ classifiers. Furthermore, we propose a combination of speaker-specific information from different languages for crosslingual speaker identification, and it is observed that the combination feature gives better performance in all the crosslingual speaker identification experiments.

Ontology is a formal, explicit specification of a shared conceptualization. Ontology provides domain vocabulary, domain knowledge, common understanding, shareability, information interoperability, reusability, concept hierarchy, and relationships that support semantic information retrieval. Ontology improves performance of the system by addressing interoperability issues due to semantic and syntactic heterogeneity. Vast numbers of application domain experts are using ontologies in diverse applications. Use of effective and efficient ontology storage system results improved performance in applications and enables semantic information retrieval. Many prominent researchers and software agencies have proposed and developed several ontology storage methods and tools with various features. The choice of a specific storage model/tool always depend on the specific purpose of the application and the nature of features that are available in the storage model/tool to be utilized in the specific applications. The familiarity of various ontology storage models and tools with the respective features helps user to choose an appropriate storage structure aiming at high-performance applications. The current research work is a comprehensively authentic study carryout out on various ontology storage models and tools with their respective features, which are very essential for optimum performance.

Extreme Learning Machine provides very competitive performance to other related classical predictive models for solving problems such as regression, clustering, and classification. An ELM possesses the advantage of faster computational time in both training and testing. However, one of the main challenges of an ELM is the selection of the optimal number of hidden nodes. This paper presents a new approach to node selection of an ELM based on a 1-norm support vector machine (SVM). In this method, the targets of SVM y i ∈{+1, –1} are derived using the mean or median of ELM training errors as a threshold for separating the training data, which are projected to SVM dimensions. We present an integrated architecture that exploits the sparseness in solution of SVM to prune out the inactive hidden nodes in ELM. Several experiments are conducted on real-world benchmark datasets, and the results attained attest to the efficiency of the proposed method.

In this paper, a cuckoo search (CS) algorithm using elite opposition-based strategy is proposed. The opposite solution of the elite individual in the population is generated by an opposition-based strategy in the proposed algorithm and form an opposite search space by constructing the opposite population that locates inside the dynamic search boundaries, then, the search space of the algorithm is guided to approximate the space in which the global optimum is included by simultaneously evaluating the current population and the opposite one. This approach is helpful to obtain a tradeoff between the exploration and exploitation ability of CS. In order to enhance the local searching ability, local neighborhood search strategy is also applied in this proposed algorithm. The experiments were conducted on 14 classic benchmark functions and 28 more complex functions from the IEEE CEC’2013 competition, and the experimental results, compared with five other meta-heuristic algorithms and four improved cuckoo search algorithms, show that the proposed algorithm is much better than the compared ones at not only the accuracy of solutions but also for the convergence speed.

In this paper, the problem of automatic data clustering is treated as the searching of optimal number of clusters so that the obtained partitions should be optimized. The automatic data clustering technique utilizes a recently developed parameter adaptive harmony search (PAHS) as an underlying optimization strategy. It uses real-coded variable length harmony vector, which is able to detect the number of clusters automatically. The newly developed concepts regarding “threshold setting” and “cutoff” are used to refine the optimization strategy. The assignment of data points to different cluster centers is done based on the newly developed weighted Euclidean distance instead of Euclidean distance. The developed approach is able to detect any type of cluster irrespective of their geometric shape. It is compared with four well-established clustering techniques. It is further applied for automatic segmentation of grayscale and color images, and its performance is compared with other existing techniques. For real-life datasets, statistical analysis is done. The technique shows its effectiveness and the usefulness.

In this paper, we will present a novel and the first approach to transforming written Arabic numeral word format into a string of digits according to proposed Arabic numeral rules. In this work, we relied mainly on the Arabic stemmer to determine the digit number, where we proposed a set of stems for the main numbers. The proposed approach covers all problems associated with Arabic numeral texts, including different numeral word shapes and unordered sentences. That is, a textual number could be written in more than one format; moreover, we deal with Arabic accents. In addition, we propose to deal with simple spelling errors and using similarities to find the nearest correct word. The proposed approach covers all possible text numbers in the range of one to millions. Tests have been executed using a possible case from context numeral in the range above, and the experimental results have demonstrated that the proposed approach is promising in the sense that it provides an efficient and error-free transformation. Finally, we built an Android mobile-based application that uses the proposed approach to transform between three different formats: digits, numeral texts, and Abacus.

A project manager balances the resource allocation using resource leveling algorithms after assigning resources to project activities. However, resource leveling does not ensure optimized allocation of resources. Furthermore, the duration and cost of a project may increase after leveling resources. The objectives of resource allocation optimization used in our research are to (i) increase resource utilization, (ii) decrease project cost, and (iii) decrease project duration. We implemented three search-based software engineering algorithms, i.e. multiobjective genetic algorithm, multiobjective particle swarm algorithm (MOPSO), and elicit nondominated sorting evolutionary strategy. Twelve experiments to optimize the resource allocation are performed on a published case study. The experimental results are analyzed and compared in the form of Pareto fronts, average Pareto fronts, percent increase in resource utilization, percent decrease in project cost, and percent decrease in project duration. The experimental results show that MOPSO is the best technique for resource optimization because after optimization with MOPSO, resource utilization is increased and the project cost and duration are reduced.