Neural Network Controller Application on a Visual based Object Tracking and Following Robot

Navigation is the main issue for autonomous mobile robot due to its mobility in an unstructured environment. The autonomous object tracking and following robot has been applied in many places such as transport robot in industry and hospital, and as an entertainment robot. This kind of image processing based navigation requires more resources for computational time, however microcontroller currently applied to a robot has limited memory. Therefore, effective image processing from a vision sensor and obstacle avoidances from distance sensors need to be processed efficiently. The application of neural network can be an alternative to get a faster trajectory generation. This paper proposes a simple image processing and combines image processing result with distance information to the obstacles from distance sensors. The combination is conducted by the neural network to get the effective control input for robot motion in navigating through its assigned environment. The robot is deployed in three different environmental setting to show the effectiveness of the proposed method. The experimental results show that the robot can navigate itself effectively within reasonable time periods.

Review of Optimization Techniques for Sizing Renewable Energy Systems

The growing evidence of the global warning phenomena and the rapid depletion of fossil fuels have drawn the world attention to the exploitation of renewable energy sources (RES). However standalone RES have been proven to be very expensive and unreliable in nature owing to the stochastic nature of the energy sources. Hybrid energy system is an excellent solution for electrification of areas where the grid extension is difficult and not economical. One of the main attribute of hybridising is to be able to optimally size each RES including storages with the aim of minimizing operation costs while efficiently and reliably responding to load demand. Hybrid RES emerges as a trend born out of the need to fully utilize and solve problems associated with the reliability of RES. This paper present a review of techniques used in recent optimal sizing of hybrid RES. It discusses several methodologies and criteria for optimization of hybrid RES. The recent trend in optimization in the field of hybrid RES shows that bio-inspired techniques may provide good optimization of system without extensive long weather data.

Coronary Heart Disease Interpretation Based on Deep Neural Network

Coronary heart disease (CHD) population increases every year with a significant number of deaths. Moreover, the mortality from coronary heart disease gets the highest prevalence in Indonesia at 1.5 percent. The misdiagnosis of coronary heart disease is a crucial fundamental that is the major factor that caused death. To prevent misdiagnosis of CHD, an intelligent system has been designed. This paper proposed a simulation which can be used to diagnose the coronary heart disease in better performance than the traditional diagnostic methods. Some researches have developed a system using conventional neural network or other machine learning algorithm, but the results are not a good performance. Based on a conventional neural network, deeper neural network (DNN) is proposed to our model in this work. As known as, the neural network is a supervised learning algorithm that good in the classification task. In DNN model, the implementation of binary classification was implemented to diagnose CHD present (representative “1”) or CHD absent (representative “0”). To help performance analysis using the UCI machine learning repository heart disease dataset, ROC Curve and its confusion matrix were implemented in this work. The overall predictive accuracy, sensitivity, and specificity acquired was 96%, 99%, 92%, respectively.

Optimization Distance Learning Computer of Network

Implementation Distance Learning (DL) lecture at High School of Police Science namely Sekolah Tinggi Ilmu Kepolisian (STIK-PTIK) consisted of 32 Polisi Daerah (POLDA) in lecture Distance Learning (DL) throughout Indonesia. System bandwidth management using the method of simple queue, the simple queue is lacking both in bandwidth allocation. Optimization against computer networks in improving Quality of Service (QoS) using the method Per Connection Queue (PCQ) Queue Tree with four classes to model. Scale model of a priority bandwidth specifically as a model of optimization of computer networks with an average percentage of delay of 6.01%, packet loss decreased 0.26%, jitter of 13.56% and increased throughput became of 9.5%. The research is supported by the level of satisfaction by CSI towards PJJ / DL students, the methods of customer satisfaction index with the service quality (Servqual) questionnaire as against with levels of satisfaction the use of DL student participants with the result satisfaction levels of 74%. Keywords: hierarchical token bucket, per connection queue, queue tree, QoS parameters.

Deep Neural Network Structure to Improve Individual Performance based Author Classification

This paper proposed an improved method for author name disambiguation problem, both homonym and synonym. The data prepared is the distance data of each pair of author’s attributes, Levenshtein distance are used. Using Deep Neural Networks, we found large gains on performance. The result shows that level of accuracy is 99.6% with a low number of hidden layers

Technical Foundations of GIS for the planning and management of the educational sector in the city of Nasiriyah

The local government of the city of Nasiriyah has always appealed for a number of problems which it described as the difficult problems facing the reality of education in the city of Nasiriyah in Iraq, pointing to a large shortage of school buildings and educational staff and overcrowded classrooms. In order to improve the status of education in general and in the city of Nasiriyah in particular, the researcher presented in this paper a number of methods and modern techniques, which are considered as the basis of the context of the unit of GIS in the Directorate of Education Dhi-Qar to study the current reality of the educational institution and to identify the foci of these problems to control and solve them in an effective and rapid manner. In this research, the programming, analytical and statistical methods of the ArcGIS program and Python programming language were used. The researcher concluded that following these principles leads to locating the focal points of the problems of the educational sector in the city of Nasiriyah, located in the Eastern Ring and the southeast of the city. Several recommendations have been made in this research to solve these problems.

Identification of Significant Proteins Associated with Diabetes Mellitus Using Network Analysis of Protein-Protein Interactions

Computation approach to identify significance of proteins related with disease was proposed as one of the solutions from the problem of experimental method application which is generally high cost and time consuming. The case of study was conducted on Diabetes Melitus (DM) type 2 diseases. Identification of significant proteins was conducted by constructing protein-protein interactions network and then analyzing the network topology. Dataset was obtained from Online Mendelian Inheritance in Man (OMIM) and Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) which provided protein data related with a disease and Protein-Protein Interaction (PPI), respectively. The results of topology analysis towards Protein-Protein Interaction (PPI) showed that there were 21 significant protein associated with DM where INS as a network center protein and AKTI, TCF7L2, KCNJ11, PPARG, GCG, INSR, IAPP, SOCS3 were proteins that had close relation directly with INS.

Classification Method of Hand Gestures Based on Support Vector Machine

This paper presents the EMG signal classification based on PCA and SVM method. The data is acquired from the 5 subjects and each subject perform 7 hand gestures includes the tripod, power, precision closed, finger point, mouse, hand open, and hand close. Each gesture is repeated 10 times (5 data as training data and the 5 remaining data as testing data). Each of training and testing data are processed using 16 features extraction in time–domain and reduced using principal component analysis (PCA) to obtain new set of features. Features classification using support vector machine classify new set of features from each subject result 85% - 89% percentage of training classification. Training data classification is tested using testing data of EMG signals and giving accuracy reach 80% - 86%.

Real-Time Lighting Control System for Smart Home Applications

The use of pervasive computing in the context of home automation equipment will greatly facilitate life. Several building still use manual switch to turn on or turn off the lighting system. It becomes ineffective if the house has a lot of lights, due to it sometimes forget to turn off. Hence, the real-time control system for automatic lighting processing is desirable. An automatic control system will allow to control the illumination and it will decrease the energy costs. In this paper, the Fuzzy logic system-based Mamdani style is used to adjust the intensity of the lights. Based on simple algorithm the controller board is working in a real-time condition. As a result found, the implementation is successfully to control the lighting system with good performance. Thus, the fuzzy system can be built smart home concept that facilitate the human life.

Improving Individual Based Author Classification Using Deep Neural Network

This paper proposed an improved method for author name disambiguation problem, both homonym and synonym. The data prepared is the distance data of each pair of author’s attributes, Levenshtein distance are used. Using Deep Neural Networks, we found large gains on performance. The result shows that level of accuracy is 99.6% with a low number of hidden layers