People with visual impairment are the second largest affected category limited access to assistive products. A complete, portable, and affordable smart assistant for helping visually impaired people navigate indoors, outdoors, interact environment is presented in this paper. The prototype of consists a cane central unit; communication between user carried out through voice messages, making system suitable any user, regardless their IT skills. equipped GPS, electronic compass, Wi-Fi,...

The core of this work consists in an explicit and complete implementation embedded system for temperature control. is developed around Arduino board uses a fuzzy logic as controller. literature abounds various solutions referring to design, implementation, and/or simulations control systems, controllers, their combinations. This offers novice designers platform understanding, experimenting learning the main concepts both closed loop system, operating real environment. major achievements can...

Nowadays, to identify potentially dangerous situations created by pedestrians, performing video surveillance systems are more than necessary. The application presented in this paper is designed automatically detect individual pedestrians and analyze characterize their behavior. Based on result, an informed security decision can be made. implementation uses YOLO (You-Only-Look-Once) [1] people a real-time stream evaluates characterizes movement, based data collected from the current frame...

Many zones, especially wild areas need surveillance systems in order to protect them against potential destroying actions. An attractive solution for automatic consists audio based systems, which present some advantages compared with video or mixt systems. This paper proposes an alarming events detection system two levels. On the first level detects only if event is a dangerous one (alarm on) normal off). second level, it case, identifies exactly nature of four classes: chainsaw, gunshot,...

This paper describes a complex software instrument, able to perform failure analysis and provide an accurate prediction of manufacturing failures. The novelty consists the association between basic quality management tool (Pareto diagram) computational intelligence techniques (neural networks). data is obtained from electronic chip process, used as input for both Pareto diagram neural network. Software instruments which using tools lack part. On other hand, networks are widely prediction....

The design and verification of complex electronic systems, especially the analog mixed-signal ones, prove to be extremely time consuming tasks, if only circuit-level simulations are involved. A significant amount can saved a cost effective solution is used for extensive analysis system, under all conceivable conditions. This paper proposes data-driven method build fast evaluate, but also accurate metamodels capable generating not-yet simulated waveforms as function different combinations...

License plate recognition is a mandatory task in evolved traffic management systems, and automating this process reduces human errors or oversights. This paper describes the development of an automatic system that uses license to send real-time notifications drivers. The combines two convolutional neural network architectures perform accurate single multiple detection recognition. plates are first detected from acquired image, then characters digits extracted, owner identified database....

To fully analyze the time response of a complex system, in order to discover its critical operation points, output waveform (under all conceivable conditions) needs be generated. Using conventional methods as physical experiments or detailed simulations can prohibitive from resources point view (time, equipment). The challenge is generate by numerous samples function different operating conditions described set parameters. In this paper, we propose fast evaluate, but also accurate model that...

This paper analyses the influence of quantization audio signals on Time Encoding Signal Processing and Recognition S-matrix, in order to detect classify intruders wildlife areas. The intruder classification is performed with multilayer feed-forward neural networks. databases involved this work consist 640 waveforms originated from 4 different types sources. experimental results proves that proposed based detection framework, overall correct rates remain very high even if number bits used for...

In the preprocessing stage that prepares data to be used in training a machine learning model, available datasets are usually highly unbalanced. This paper analyzes and compares three methods for solving issue of imbalanced datasets, using from electronic payment transactions, where fraudulent operations represent minority class. The Python implementation proves SMOTE with Tomek Links technique provides best results, based on analyzing F1-scores, Precision-Recall ROC curves.

The present paper presents the implementation of an intrusion detection system prototype which is based on plastic optical fiber. can serve in monitoring different types areas starting from small rooms up to dimensions bigger grounds and territories. Unlike commercially available systems, proposed alternative invisible for intruders, has a simple functioning principle scalable. basic that underlines if pressure applied fiber, light flux pass through fiber will be reduced, variation being...

The paper proposes an automated optical inspection system for detection of PCB mounting defects. reference image was processed using HSV color space and thresholding. mask created the colors mounted components - black, gray brown. can identify multiple faults on test in a single cycle, thus increasing efficiency process. is suitable both bare populated PCBs.

Signal approximation is a matter of great interest, as working with complete time-sampled signals requires memory and computational resources. In order to diminish these requirements, signal compression methods are widely used. The paper proposes method, using genetic algorithm that guides the wavelet decomposition process, by providing specific information, such as: mother wavelet, number selected coefficients, level. tradeoff between quality its complexity addressed in objective function...

This paper explains why and how the very same circuit (a transistor in CS connection) is able to perform multiple applications depending on use of different operating regions. Families transfer characteristics are used deduce possible applications. In this way, emphasis transferred understanding intrinsic operation a circuit, which opens perspective for applications, as opposed traditional approach, where placed

The interpretation of data gathered from dam monitoring directly influences the detection abnormal behaviors. Using previously recorded data, predictive models can be developed, so that signs a possible failure are detected as early possible. paper presents multi-step ahead model to generate values for horizontal displacement dam, using previous displacement, water level and temperature. is based on an autoregressive neural network was trained tested historical data. results show good...

The new generation of students comes with demands regarding the way they are taught, from kindergarten to university. In engineering education, a continuous effort needs be made, in order provide both theoretical knowledge, available for remote learning, as well practical skills, developed through face-to-face activities. For electronic devices and circuits lab, transition Education 4.0 was made gradually. knowledge is online, studied preparation while skills weekly lab sessions. six working...

This paper presents the design and analysis of a sub-1 dB noise figure CMOS single-ended Low Noise Amplifier (LNA), in 0.35 μm technology. The LNA operates at 2.1 GHz, addressing Wideband Code Division Multiple Access (WCDMA) standard. Simulations prove that gain input return loss satisfy requirements this standard, with value way less than specified. Temperature, voltage supply process variation sweeps robustness design.

Academic teaching performance is recognized as a top instrument in reducing student dropout. This paper proposes fuzzy logic based expert system which helps the teachers perform their own evaluation of academic and research activities, order to match them requirements for each study cycle (introductory, specialization, master). The developed also offers way monitoring progress teachers’ performances provide additional explanations about aspects still need improvement. was tested using...

This paper proposes a hybrid system for evaluating the dropout risk students enrolled in first year of tertiary education, based on data collected from an engineering faculty Romania. The determines profile student and matches it against series predefined values certain indicators, checking whether student’s is at alarming level. used developing was by means anonymous online forms, 1st students, during semester. So far, tested using artificial only, but currently applied Faculty real results...

Fault detection and diagnosis are fundamental steps in maintaining high quality standards electronic circuits mass production, while keeping the costs at a minimal level. This paper proposes an automated method for designing fuzzy logic system fault of circuits. Data from real circuit its simulated version is acquired processed, so that system's rules can be established. The model faulty functioning circuit, as well fault-free state. activation degree each rule will indicate component....