Jan Matuszewski

Dr. Jan Matuszewski


Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 December 2016 in Annual of Navigation
Verified
Journal Article
Open access
PDF file
SKStanislaw Konatowski
Stanislaw Konatowski
Piotr KaniewskiPiotr Kaniewski
Piotr Kaniewski
JMJan Matuszewski
Jan Matuszewski
Several types of nonlinear filters (EKF – extended Kalman filter, UKF – unscented Kalman filter, PF – particle filter) are widely used for location estimation and their algorithms are described in this paper. In the article filtering accuracy for non-linear form of measurement equation is presented. The results of complex simulations that compare the quality of estimation of analyzed non-linear filters for complex non-linearities of state vector are presented. The moves of maneuvering object are described in two-dimensional Cartesian coordinates and the measurements are described in the polar coordinate system. The object dynamics is characterized by acceleration described by the univariate non-stationary growth model (UNGM) function. The filtering accuracy was evaluated not only by the root-mean-square errors (RMSE) but also by statistical testing of innovations through the expected value test, the whiteness test and the WSSR (weighted sum squared residual) test as well. The comparison of filtering quality was done in the MATLAB environment. The presented results provide a basis for designing more accurate algorithms for object location estimation.

9 December 2021 in Sensors
Verified
Journal Article
Open access
PDF file
JMJan Matuszewski
Jan Matuszewski
DPDymitr PIETROW
Dymitr PIETROW
With the increasing complexity of the electromagnetic environment and continuous development of radar technology we can expect a large number of modern radars using agile waveforms to appear on the battlefield in the near future. Effectively identifying these radar signals in electronic warfare systems only by relying on traditional recognition models poses a serious challenge. In response to the above problem, this paper proposes a recognition method of emitted radar signals with agile waveforms based on the convolutional neural network (CNN). These signals are measured in the electronic recognition receivers and processed into digital data, after which they undergo recognition. The implementation of this system is presented in a simulation environment with the help of a signal generator that has the ability to make changes in signal signatures earlier recognized and written in the emitter database. This article contains a description of the software’s components, learning subsystem and signal generator. The problem of teaching neural networks with the use of the graphics processing units and the way of choosing the learning coefficients are also outlined. The correctness of the CNN operation was tested using a simulation environment that verified the operation’s effectiveness in a noisy environment and in conditions where many radar signals that interfere with each other are present. The effectiveness results of the applied solutions and the possibilities of developing the method of learning and processing algorithms are presented by means of tables and appropriate figures. The experimental results demonstrate that the proposed method can effectively solve the problem of recognizing raw radar signals with agile time waveforms, and achieve correct probability of recognition at the level of 92–99%.

11 February 2020
Verified
Conference Paper
Open access
PDF file
JMJan Matuszewski
Jan Matuszewski
ARAdam Rajkowski
Adam Rajkowski
The article presents a way of using machine learning algorithms to recognize objects in images. To implement this task, an artificial neural network was used, which has a high adaptability and allows work with a very large set of input data. The neural network was described using a program written in the MATLAB simulation environment. The basic problem faced by the designer of objects recognition is to collect a sufficient training set of images to achieve the high probability of correct recognition. The set of learning patterns in the artificial neural networks may contain from several dozen thousands to one million training samples. In this article at the beginning the neural network was pre-trained trained based on the images included in the publicly available CIFAR 100 database, which are characterized by a small size of 32x32 pixels. It contains 70 000 images assigned to 10 basic categories. Then the author's database, consisting from 1000 pedestrians, cars and road signs was used. The article contains a description of applied algorithm, method of supervised learning and correction of weight coefficients, selection of activation function and operation on max pooling filter. The results of proposed solution are presented in the form of screenshots from calculations and in figures depicting results of recognized objects. Attention was also paid to the impact of used database for learning the network on the speed of calculations and recognition efficiency. The proper selection of number and types of layers, number of neurons, activation function and the value of the learning factor is very important in designing the neural network in application to objects recognition contained in the images. The problems occurring in the process of learning the neural networks and suggestions for their further improvement are also presented.

1 December 2017 in Annual of Navigation
Verified
Journal Article
Open access
JMJan Matuszewski
Jan Matuszewski
WGWojciech Grzywacz
Wojciech Grzywacz
The article presents navigation system project operating on the principle scene matching area correlation (SMAC), using a digital camera, an MEMS e-compass sensor and an ultrasonic ranging module. Systems of this type are used as a component of advanced integrated navigation systems in view of its autonomy and capability of localizing aircrafts with high accuracy and precision. Steering and display of information are implemented using a computer application designed in Matlab programming environment. The object’s location is fixed, using discrete cross-correlation function through matching of the registered terrain image to digital orthophotomap. The article describes operations directly related to digital image processing, its implementation methods, a structural system design with explanations of each of the functional elements and presents devices used to build a complete integrated measurement unit model. It was used for the effectiveness measurement of determining the location of an object depending on the changes of angle and height of the flight as well as the luminance and noise level in a registered image. The measurements methodology was described which also includes an analysis of the results, an effectiveness evaluation and potential development directions of the designed system.

11 February 2020
Verified
Conference Paper
Open access
PDF file
JMJan Matuszewski
Jan Matuszewski
MZMarcin Zając
Marcin Zając
The article presents a method of recognizing alphanumeric characters located in the image, based on a previously created database of patterns using neural networks. For this purpose the convolutional networks were used, which independently search for features that allow to distinguish characters in the image. A larger number of convolution layers allows us to recognize a greater number of features and thus to increase the probability of correctly recognized characters. The main purpose of the paper is to present software that recognizes the alphanumeric characters in images and to investigate the impact of the size of this database on the program's speed and character recognition efficiency. This software can also be used in more complex structures, such as automatic translators or as a computer reader. The calculation of the first program that recognizes single character and the second program that reads all the text from the image have been made in the MATLAB environment. The paper describes the components of this software, such as the learning subsystem and the character recognition subsystem. The results of the program were presented in the form of screenshots showing the results of the learning process and character recognition process. The speed of the software and the effectiveness of recognizing alphanumeric characters using the artificial neural networks and maximally stable extremal regions (MSER) algorithm are presented in the table and figures. Attention was also paid to the impact of the size of the database used to learn the network on the speed of calculations and recognition efficiency.

5 November 2019 in PRZEGLĄD ELEKTROTECHNICZNY
Verified
Journal Article
Closed access
JMJan Matuszewski
Jan Matuszewski
Wydawnictwo SIGMA-NOT wydaje czasopisma fachowe informuj?ce swoich czytelnik?w o najnowszych osi?gni?ciach naukowych i nowoczesnych rozwi?zaniach technicznych w Polsce i na ?wiecie, popularyzuje problemy techniczne oraz poszerza wiedz? i kultur? techniczn?.