Intelligent Computing by Valentyn N Sichkar
IEEE, 2019
The paper is devoted to the research of two approaches for global path planning for mobile robots... more The paper is devoted to the research of two approaches for global path planning for mobile robots, based on Q-Learning and Sarsa algorithms. The study has been done with different adjustments of two algorithms that made it possible to learn faster. The implementation of two Reinforcement Learning algorithms showed differences in learning time and the methods of building path to avoid obstacles and to reach a destination point. The analysis of obtained results made it possible to select optimal parameters of the considered algorithms for the tested environments. Experiments were performed in virtual environments where algorithms learned which steps to choose in order to get a maximum payoff and reach the goal avoiding obstacles.
Bookmarks Related papers MentionsView impact
St. Petersburg State Polytechnical University Journal. Computer Science. Telecommunications and Control Systems, Jun 1, 2018
Developing systems for intelligent navigation is one of the major problems in world of modern rob... more Developing systems for intelligent navigation is one of the major problems in world of modern robotics. This problem is particularly urgent when the environment is unknown. It means that a mobile robot meeting unpredictable obstacles on its way and has to react according to the current situation fast and in real time. That is why developing such a system is always a big challenge. This paper studies different techniques for storing and using the knowledge in order to avoid collisions with obstacles. Most attention is paid for developing two types of Knowledge Bases to help the mobile robot to avoid possible collisions and continue its way. A comparison analysis is provided for these two different types of Knowledge Bases. The advantages and disadvantages were analyzed and described.
Bookmarks Related papers MentionsView impact
The main objectives of the course project are: to learn the types of Knowledge Base; to build a r... more The main objectives of the course project are: to learn the types of Knowledge Base; to build a real system in Arduino; to create scenarios and diagrams for different systems with descriptions; to build Neural Networks and explanations of the Rules; to create applications in C# which communicates with Serial Port; to create Alarm-Warning system based on Arduino and WF C# application; to create full description of the Alarm-Warning system.
Bookmarks Related papers MentionsView impact
In modern world of automation systems, movement of objects whether inside factory or outside fact... more In modern world of automation systems, movement of objects whether inside factory or outside factory for loading on ship, always has to be safety. Objects can be large and require a system that controls states around them. Monitoring can be done by people around the objects, but in this way, there are restrictions. The most significant restriction is related to range of operability. Also, there are delays in receiving commands, and everything can be confusing by using communication of several people. In order to overcome these restrictions, the system should provide a user-friendly interface which can show immediately and in real time the states from any side of the object to one person. Therefore, ultrasonic sensors based automation system to control conditions remotely from anywhere around the object, anytime is needed. The proposed Ultrasonic Control System consists of a microcontroller hardware side and an application which forms the software side of the project. The system will work wirelessly via Bluetooth connection. The microcontroller will communicate with ultrasonic sensors to achieve the objective.
Bookmarks Related papers MentionsView impact
The main goal of this work is to get the information from the Knowledge Base (KB) about the curre... more The main goal of this work is to get the information from the Knowledge Base (KB) about the current statement for each of five sensors separately, showing the results by lighting the LEDs according to the information from the KB. There are two LEDs situated near each sensor – in red and green colors. If the information from KB is " Alarm " – red LED has to be lighted, if the information is " Warning " – green LED has to be lighted. This system can be used for auto-parking by controlling the safety distances in different directions, and also can be a part of intelligent drive assistant system of the car. In general, this study helps better to understand hardware and software representation of this idea and obtain new ideas for future improvements.
Bookmarks Related papers MentionsView impact
Image Processing by Valentyn N Sichkar
Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 2020
Article in English For citation: Sichkar V.N., Kolyubin S.A. Real time detection and classificati... more Article in English For citation: Sichkar V.N., Kolyubin S.A. Real time detection and classification of traffic signs based on YOLO version 3 algorithm. Abstract The issue of effective detection and classification of various traffic signs is studied. The two-stage method is proposed for creation of holistic model with end-to-end solution. The first stage includes implementation of effective localization of traffic signs by YOLO version 3 algorithm (You Only Look Once). At the first stage the traffic signs are grouped into four categories according to their shapes. At the second stage, an accurate classification of the located traffic signs is performed into one of the forty-three predefined categories. The second stage is based on another model with one convolutional neural layer. The model for detection of traffic signs was trained on German Traffic Sign Detection Benchmark (GTSDB) with 630 and 111 RGB images for training and validation, respectively. Сlassification model was trained on German Traffic Sign Recognition Benchmark (GTSRB) with 66000 RGB images on pure "numpy" library with 19 × 19 dimension of convolutional layer filters and reached 0.868 accuracy on testing dataset. The experimental results illustrated that the training of the first model deep network with only four categories for location of traffic signs produced high mAP (mean Average Precision) accuracy reaching 97.22 %. Additional convolutional layer of the second model applied for final classification creates efficient entire system. Experiments on processing video files demonstrated frames per second (FMS) between thirty-six and sixty-one that makes the system feasible for real time applications. The frames per second depended on the number of traffic signs to be detected and classified in every single frame in the range from six to one.
Bookmarks Related papers MentionsView impact
Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 2019
The paper presents the study of an effective classification method for traffic signs on the basis o... more The paper presents the study of an effective classification method for traffic signs on the basis of a convolutional neural network with various dimension filters. Every model of convolutional neural network has the same architecture but different dimension of filters for convolutional layer. The studied dimensions of the convolution layer filters are: 3 × 3, 5 × 5, 9 × 9, 13 × 13, 15 × 15, 19 × 19, 23 × 23, 25 × 25 and 31 ×31. In each experiment, the input image is convolved with the filters of certain dimension and with certain processing depth of image borders, which depends directly on the dimension of the filters and varies from 1 to 15 pixels. Performances of the proposed methods are evaluated with German Traffic Sign Benchmarks (GTSRB). Images from this dataset were reduced to 32 × 32 pixels in dimension. The whole dataset was divided into three subsets: training, validation and testing. The effect of the dimension of the convolutional layer filters on the extracted feature maps is analyzed in accordance with the classification accuracy and the average processing time. The testing dataset contains 12000 images that do not participate in convolutional neural network training. The experiment results have demonstrated that every model shows high testing accuracy of more than 82%. The models with filter dimensions of 9 × 9, 15 × 15 and 19 × 19 achieve top three with the best results on classification accuracy equal to 86.4 %, 86 % and 86.8 %, respectively. The models with filter dimensions of 5 × 5, 3 × 3 and 13 × 13 achieve top three with the best results on the average processing time equal to 0.001879, 0.002046 and 0.002364 seconds, respectively. The usage of convolutional layer filter with middle dimension has shown not only the high classification accuracy of more than 86 %, but also the fast classification rate, that enables these models to be used in real-time applications.
Bookmarks Related papers MentionsView impact
2018 International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM), 2018
An access control system by fingerprints is one of the most widely spread biometric techniques li... more An access control system by fingerprints is one of the most widely spread biometric techniques linked with ensuring the security of the enterprise. The uniqueness of fingerprints is something staying unchanged during the whole life. For example, such features as the face shape, manner of walking may change over time time but fingerprints will not. However, the fingerprint itself can give an image of a bad quality because of the state of the skin, bad impression, dirty surface of the finger, etc. That's why, the methods for image enhancement should be implemented to get more accurate results. This study discusses reliable techniques and propose a reliable algorithm of the fingerprint image enhancement for the access control system to be applied as a part of the security system in the company.
Bookmarks Related papers MentionsView impact
Information Security by Valentyn N Sichkar
When a company starts to build enterprise network, the information technology department always f... more When a company starts to build enterprise network, the information technology department always faces the question about which protocol to choose for building a Virtual Private Network (VPN). The answer is not obvious because each protocol has its own pros and cons. Priority of the selection of the protocol to build the VPN depends on specific criteria. These criteria can be the required type of access for users, the level of the network security, the level of data security, necessity of scalability for the network in future, etc. At the moment the most reliable protocols are considered to be Internet Protocol Security (IPSec) and Secure Socket Layer (SSL). This paper discusses and analysis the demand of using these protocols separately or simultaneously.
Bookmarks Related papers MentionsView impact
In a world where digital communications play a key role, encryption is an elementary security rul... more In a world where digital communications play a key role, encryption is an elementary security rule that prevents a huge number of problems. When we send an e-mail, it passes through dozens of intermediate nodes. And each of the nodes can have many vulnerabilities and security holes which can be used by anyone. For many years, up to the 70s of the last century, a single key was used for encryption and decryption. This single key could be a key phrase, a combination of the details of the encryption machine, etc., which should have been known, and to the one who encrypts the message, and the one who should read it. This is the so-called symmetric encryption. The main problem of this method is the exchange of keys. Public keys were developed to solve the problem of secret key exchange. GPG encryption method is an excellent tool for encrypting e-mail and digital materials. It uses a key pair to send messages. One key is open (public key) and can be transferred to anyone. The other key is secret (private key) and is stored only by the owner. The sender encrypts the message with the own secret key and the help of public key of the recipient and now the encrypted message can be decrypted only with using of the receiver's private key and with the public key of the sender. This method solves the secret key exchange problem inherent in symmetric ciphers. There is no need to agree on a key between the sender and the recipient. All that is required before the start of secret exchange is the exchange of public keys between sender and receiver. In addition, one public key can be used by all correspondents of the recipient.
Bookmarks Related papers MentionsView impact
Conference Presentations by Valentyn N Sichkar
Implementing Convolutional Neural Network on pure "numpy" library for Image Classification on the... more Implementing Convolutional Neural Network on pure "numpy" library for Image Classification on the basis of CIFAR-10 dataset and MNIST handwritten digits.
Bookmarks Related papers MentionsView impact
Implementing Convolutional Neural Network (CNN) for Classification of Traffic Signs. CNN trained ... more Implementing Convolutional Neural Network (CNN) for Classification of Traffic Signs. CNN trained on German Traffic Sign Benchmarks (GTSRB). Results can be tested online on the web, where trained CNN were deployed.
Bookmarks Related papers MentionsView impact
Uploads
Intelligent Computing by Valentyn N Sichkar
Image Processing by Valentyn N Sichkar
Information Security by Valentyn N Sichkar
Conference Presentations by Valentyn N Sichkar