Studies in Systems, Decision and Control 33. Aleksander Nawrat Karol Jędrasiak Editors. Innovative Simulation Systems

Transcription

1 Studies in Systems, Decision and Control 33 Aleksander Nawrat Karol Jędrasiak Editors Innovative Simulation Systems

2 Studies in Systems, Decision and Control Volume 33 Series editor Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland

3 About this Series The series Studies in Systems, Decision and Control (SSDC) covers both new developments and advances, as well as the state of the art, in the various areas of broadly perceived systems, decision making and control- quickly, up to date and with a high quality. The intent is to cover the theory, applications, and perspectives on the state of the art and future developments relevant to systems, decision making, control, complex processes and related areas, as embedded in the fields of engineering, computer science, physics, economics, social and life sciences, as well as the paradigms and methodologies behind them. The series contains monographs, textbooks, lecture notes and edited volumes in systems, decision making and control spanning the areas of Cyber-Physical Systems, Autonomous Systems, Sensor Networks, Control Systems, Energy Systems, Automotive Systems, Biological Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems, Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems, Robotics, Social Systems, Economic Systems and other. Of particular value to both the contributors and the readership are the short publication timeframe and the world-wide distribution and exposure which enable both a wide and rapid dissemination of research output. More information about this series at

4 Aleksander Nawrat Karol Jędrasiak Editors Innovative Simulation Systems 123

5 Editors Aleksander Nawrat Institute of Automatic Control Silesian University of Technology Gliwice Poland Karol Jędrasiak Institute of Automatic Control Silesian University of Technology Gliwice Poland ISSN ISSN (electronic) Studies in Systems, Decision and Control ISBN ISBN (ebook) DOI / Library of Congress Control Number: Springer Cham Heidelberg New York Dordrecht London Springer International Publishing Switzerland 2016 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. Printed on acid-free paper Springer International Publishing AG Switzerland is part of Springer Science+Business Media (

6 Be nice to people on your way up because you meet them on your way down. Jimmy Durante

7 Preface The book has been motivated by rapid development of technology allowing us to create innovative simulation systems designed to deliver broader content; at the same time enhancing educational opportunities and reducing the economic cost of teaching, in particular, in areas such as national defense, medicine, or sport. The concept of a simulation system is extensive and contains high-growth areas such as augmented reality or design and construction of unmanned vehicles. Augmented reality (AR) is a term used to describe the process of enriching the perception of the real world with digital elements, generated by a computer. Augmentation of reality is typically performed in real time and in a contextual conformity with the observed real world. Mainly, the information added is a multimedia extension of the senses of the observer. Many research groups are working on projects and implementation of simulation systems containing elements of AR, which are used in many fields such as navigation, rehabilitation, entertainment, and video surveillance of the city. The main aim of design and construction of UAVs is their practical application. However, their design is a lengthy process and requires intensive evaluation using both simulation and experimental data. Human perception is limited to tracking in full concentration only seven elements simultaneously. Control of unmanned vehicle and analysis of data recorded by the payload of unmanned vehicles is a significant excess of information for human and negatively impacts its analytical capabilities. Therefore, it is often necessary to apply amendments based on the actual real life experience of the application. The goal of this monograph is to provide comprehensive guidelines on the current and future trends of innovative simulation systems; in particular, their important components, such as augmented reality and unmanned vehicles. The book consists of three parts: Vision-based Information for Innovative Simulation Systems, Design, Construction and Analysis for Purpose of Innovative Systems and Design and Evaluation of Control Algorithms. Each part presents good practices, new methods, concepts of systems, and new algorithms. Presented challenges and solutions are the results of research conducted by the contributing authors. vii

8 viii Preface The monograph can be a valuable source of information for those in contact with the design and creation of innovative simulation systems as researchers, programmers, designers, engineers, and users. The book is addressed to a wide audience: academic staff, representatives of research institutions, employees of companies and government agencies as well as students and graduates of technical universities in the country and abroad. We took an effort to present multiple solutions and challenges in the fields above. The authors encourage to actively explore the following chapters of the book and, in case of questions or concerns, invite you to an open and frank discussion of these matters. In conclusion, the main features of the book are the following: 1. The book describes and evaluates the current state of knowledge in the field of innovative simulation systems. 2. Throughout the chapters there are presented current issues and concepts of systems, technology, equipment, tools, research challenges and current, past, and future applications of simulation systems. We would like to thank the authors for their contributions. Without their knowledge, experience, and commitment this monograph would have never appeared. We also thank the reviewers, editors, and staff of Springer for their support during the creation of this monograph. Gliwice April 2015 Aleksander Nawrat Karol Jędrasiak

9 Acknowledgment The work was financially supported by NCBiR grant POIG GeCONiI Upper Silesian Centre for Computational Science and Engineering. ix

10 Contents Part I Vision-Based Information for Innovative Simulation Systems Minimizing the Image Resolution in Order to Increase the Computing Speed Without Losing the Separation of the Recognised Patterns... 3 Zygmunt Kuś and Aleksander Nawrat The Method of Guaranteeing the Separation Between the Recognised Object and Background Zygmunt Kuś and Aleksander Nawrat The Method of Developing the Invariant Functions Vector for Objects Recognition from a Given Objects Set Zygmunt Kuś and Aleksander Nawrat Adjusting the Thresholds to the Recognised Pattern in Order to Improve the Separation Between the Recognised Patterns Zygmunt Kuś and Aleksander Nawrat The Concept of an Active Thermal Camouflage for Friend-Foe Identification System Paulina Wilk, Tomasz Targiel, Dawid Sobel, Jan Kwiatkowski, Karol Jędrasiak and Aleksander Nawrat Augmented Reality in UAVs Applications Paweł Iwaneczko, Karol Jędrasiak and Aleksander Nawrat xi

11 xii Contents Real Time Dense Motion Estimation Using FPGA Based Omnidirectional Video Acquisition Device Jan Kwiatkowski, Dawid Sobel, Artur Ryt, Mariusz Domżał, Karol Jędrasiak and Aleksander Nawrat GPU-Based Parameters Estimation for Anisotropic Diffusion Mariusz Domżał, Artur Ryt, Dawid Sobel, Jan Kwiatkowski, Karol Jędrasiak and Aleksander Nawrat An Evaluation of SETh The Method for Long-Term Feature Tracking Karol Jędrasiak and Aleksander Nawrat The System for Augmented Reality Motion Measurements Visualization Dawid Sobel, Jan Kwiatkowski, Artur Ryt, Mariusz Domżał, Karol Jędrasiak, Łukasz Janik and Aleksander Nawrat Real-Time Multiple Laser Points Tracking Artur Ryt, Dawid Sobel, Jan Kwiatkowski, Mariusz Domżał, Karol Jędrasiak and Aleksander Nawrat Part II Design, Construction and Analysis for Purpose of Innovative Systems Numerical and Experimental Analysis of a Truck Frame Eugeniusz Rusiński, Artur Iluk and Mariusz Stańco Designing and Implementing Elements of a Vehicle Model in the VBS2 Virtual Simulation Environment Roman Wantoch-Rekowski and Konrad Szumiec Semi-active Suspension System for 2S1 Tracked Platform in Drive Comfort Improvement Application Tomasz Nabagło, Andrzej Jurkiewicz and Janusz Kowal Experimental Mobile Robot Hardware Krzysztof Jaskot and Tomasz Łakota Conception of a Diagnostic System for Evaluating a Technique Correctness and Effectiveness of Running Krzysztof Skrzypczyk

12 Contents xiii Part III Design and Evaluation of Control Algorithms Experimental Mobile Robot Software Krzysztof Jaskot and Tomasz Łakota Application of Genetic Algorithms for Identification of Simulated Systems Martyna Ulinowicz and Janusz Narkiewicz Quadrotor Dynamics and Control for Precise Handling Przemysław Bibik, Janusz Narkiewicz, Maciej Zasuwa and Marcin Żugaj The Application of the Modified BLT Method for the Synthesis of UAV s Control System Zygmunt Kuś and Aleksander Nawrat The Modified BLT Method for Multivariable Control Systems Zygmunt Kuś and Aleksander Nawrat Analysis of Human Arm Nonlinear and Linear Mathematical Models Robert Bieda and Krzysztof Jaskot Naive Kalman Filtering for 3D Object Orientation Robert Bieda, Rafal Grygiel and Adam Galuszka Non-GPS Navigation System for Criminalistic Investigation on Mobile Robot Robert Głębocki, Paweł Kicman and Janusz Narkiewicz Path Planning Algorithms for Autonomous Mobile Platform Krzysztof Jaskot and Radomir Mysliwiec

13 Part I Vision-Based Information for Innovative Simulation Systems Progress of miniaturization and increase in resolution of widely available video cameras allowed a rapid progress of vision-based innovative simulation systems. With the advent of high-quality smartphones enabling not only the acquisition, but at the same time processing and displaying of data, it became possible practical application of the concept of augmented reality (AR). The part presents an example of Attitude Indicator Augmented Reality (AR) control for the unmanned aerial vehicles (UAV) ground control station. The article describes perspective heads-up display overlaid on the image, which is acquired from the UAV rotatable camera. Article shows the mechanisms of aircraft camera calibration with the AR artificial horizon. The whole algorithm of the render instructions is presented. Elements used in the augmented reality are as follows: 3D artificial horizon, latitude and longitude, GPS info, executed command, time to command end, percent of command accomplish, fuel and battery level, height and speed vertical scale, landing field direction arrow, unmanned vehicle marks. Another discussed application of AR is an attempt to improve existing methods of measuring the range of motion of people during rehabilitation. Range of motion measurements are used by physicians, physical therapists and veterinarians. There is a need for a solution that would increase the accuracy and repeatability of the measurements obtained. Commonly used methods based on tests using a classical goniometer are often inaccurate and depend on the individual interpretation of result by the therapist. The use of motion capture data to RoM measurements results in an accuracy of order of tens of micrometers, which is mainly due to the precision of the Motion Capture system. During the test, patient is not constrained by any device and has a total freedom of movement. It allows to carry out a medical examination not only during flexion and extension of the one limb but more limbs during dynamic movement such as walking. However, one should keep in mind that to a large extent the effectiveness of augmented reality systems depends on use for its implementation video acquisition components and tracking algorithms.

14 2 Part I: Vision-Based Information for Innovative Simulation Systems One of the main problems of image acquisition by the UAV is the need for the use of opto-electronic gimbals which tend to be expensive, and the moving parts are subject to frequent breakdowns. An alternative may be a multi objective solution devoid of moving parts. One of the solutions designed in such a way is Omnidirectional Video Acquisition Device (OVAD), a device, which was implemented using analog cameras and STM32 microcontroller, has proven its validity during testing process. The device may be found useful in many applications, both civilian and military. However, size and weight of the device made it difficult to implement OVAD on small unmanned vehicles. Proposed solution is FPGA-based OVAD, which is the result of OVAD further development. Due to the significant differences in the acquired image from thermal imaging cameras and visible light cameras it is often necessary to develop independent algorithms for filtering and tracking the objects of interest pointed by the human operator. Within the part there is presented a novel long-term object tracking method called SETh. It is an adaptive tracking by detection method which allows near real-time tracking within challenging sequences. The algorithm consists of three stages: detection, verification and learning. In order to measure the performance of the method, video data set consisting of more than a hundred videos was created and manually labeled by a human. Quality of the tracking by SETh was compared against five state-of-the-art methods. The presented method achieved results comparable and mostly exceeding the existing methods, which proved its capability for real life applications like vision-based control of UAVs. Another study was to examine the influence of image thresholding on the correctness of pattern recognition in grey scale images. The method based on moment invariants, which were the elements of feature vectors defining the features of the recognized object, was used by authors in order to recognize the objects. The article presents the influence of image thresholding with histogram equalization for exemplary images on the distribution of the distance between pattern vector and feature vector for every pixel of an image. The authors have paid great attention to the fact that proper selection of the thresholds is significant for distinguishing given object classes. The aim of the following study was to develop a procedure which guarantees the separation between the part of an image where we have the recognized object and the part of the image which corresponds to the terrain where the object moves. The authors have presented the method which uses moment invariants for creating feature vectors which define the features of the recognized object and the features of the background. The presented method is based on calculating the distance between the values of invariant functions calculated for an object and the background. On top of that, the study presents the example of recognizing the object moving in various types of a terrains. Concluding, the part includes a number of important challenges in the fields mentioned above. At the same time valuable suggestions and conclusions from authors are presented and discussed in detail.

15 Minimizing the Image Resolution in Order to Increase the Computing Speed Without Losing the Separation of the Recognised Patterns Zygmunt Kuś and Aleksander Nawrat Abstract The aim of the following study was to examine the influence of image resolution in the pattern recognition in the grey scale images. In order to recognise the pattern, the authors used the method based on moment invariants which were the elements of feature vectors defining the features of the recognised object. The paper presents the influence of image resolution for exemplary images on both: the values of moment invariants and distances between feature vectors. The authors have paid a great attention to the fact that these distances are significant for distinguishing given object classes. One can conclude from the results that for a significant decrease in resolution there occur problems in pattern recognition. It results from the influence of image resolution on the value of moment invariants and at the same time on the value of the distance between the feature vectors defining the recognised objects. In this way, the paper shows that in order to recognise objects correctly, it is necessary to retain some necessary minimum resolution. It is indispensable despite the fact that we usually aim at decreasing the amount of processed data which is on the hand crucial because of short processing times in many practical applications. It is therefore essential because of the fact that we need to guarantee short times of image processing in many practical applications. Moreover, this study presents the examples of the algorithms in use. Keywords Object recognition Moment invariants Pattern vector Image resolution Z. Kuś (&) A. Nawrat Institute of Automatic Control, Silesian University of Technology, Akademicka 16 Street, Gliwice, Poland zygmunt.kus@polsl.pl A. Nawrat aleksander.nawrat@polsl.pl Springer International Publishing Switzerland 2016 A. Nawrat and K. Jędrasiak (eds.), Innovative Simulation Systems, Studies in Systems, Decision and Control 33, DOI / _1 3

16 4 Z. Kuś and A. Nawrat 1 Introduction Image processing encounters multiple problems in the preliminary stages of image processing. One of the first stages in image processing is to decide about the image resolution. It lets us obtain a less amount of data to process; however, it should not result in deteriorating pattern recognition. There are some problems during the acquisition of the image which influence the required image resolution and consequently they impact on the quality of the pattern recognition. Firstly, each measurement in the real world is burdened with some error and can be disturbed by a some kind of noise [1 3]. When the camera works in various atmospheric conditions, then rain, snow or the reflections of the sunlight can result in an image with noise. Secondly, the amount of data which can be stored and processed is usually smaller than the amount of data acquired by the camera system [4, 5]. Therefore, on the one hand, we would like to have the highest resolution to distinguish artefacts in the image, which are the effect of disturbance, from the significant elements of the image. On the other hand, we would like to have the smallest amount of data to store and process. The following study will focus on developing a method of calculating the resolution for the acquisition of the grey scale images. Changing the resolution will be conducted in order to decrease the amount of data defining the image and it will be carried out to speed up the calculations during the recognition of the pattern (object) location in the image. Decreasing the data included in the image after decreasing the resolution must not decrease the effectiveness of pattern recognition in the image [6 8]. Many solutions of a pattern recognition methods can be found in literature [9 15] and a lot of them is based on the feature vectors which elements are moment invariants [16, 17]. This paper will study the influence of decreasing the image resolution on the values of the moment invariants which are used to recognise an object. We will analyse the images of 256 grey scale levels. 2 Formulating the Problem of the Resolution Change for Grey Scale Images The problem of the image resizing in image processing is discussed at greater length in the literature. In the body of literature, one can find various techniques for example: nearest neighbour, bilinear interpolation, or bicubic interpolation. The image resizing, either upscaling or downscaling resulting from the need for increased or decreased image size is presented in [18]. The image resizing method presented in [18] is based on the analysis of the sum of primary implicants representation of image data, as generated by a logical transform. The algorithm used in the aforementioned study emulates the nearest neighbour technique, while

17 Minimizing the Image Resolution in Order to Increase 5 subsequent variations build on this to provide more accuracy and output image comparable to the other traditional methods. The other approach to this problem is discussed in [19]. This paper presents an adaptive interpolation algorithm which is based on the Newton Polynomial in order to reduce the limitation of the traditional algorithm for image resizing. In the next paper [20], a comprehensive study and comparison of different image scaling algorithms was presented. In [20], the authors compared the major methods of image resizing. These methods comprised of nearest neighbour image scaling, bilinear image scaling, bicubic image scaling, Lanczos image scaling and modified bicubic image scaling. We will use the method based on calculating the mean value to change the image resolution. Moreover, we are going to use the moment invariants as invariant functions for pattern recognition. These moment invariants ðj 1 J 8 Þ, which are presented in (1) (3), will create the feature vector. J 1 ¼ I 1 m 2 ; J 2 ¼ I 2 00 m 4 ; J 3 ¼ I 3 00 m 5 ; J 4 ¼ I 4 00 m 5 ; 00 J 5 ¼ I 5 m ; J 6 ¼ I 6 m 7 ; J 7 ¼ I 7 00 m 4 ; J 8 ¼ I 8 00 m 5 ; 00 ð1þ I 1 ¼ M 20 þ M 02 I 2 ¼ðM 20 M 02 Þ 2 þ 4M 2 11 I 3 ¼ðM 30 3M 12 Þ 2 þð3m 21 M 03 Þ 2 I 4 ¼ðM 30 þ M 12 Þ 2 þðm 21 þ M 03 Þ 2 I 5 ¼ðM 30 3M 12 ÞðM 30 þ M 12 ÞððM 30 þ M 12 Þ 2 3ðM 21 þ M 03 Þ 2 Þ þð3m 21 M 03 ÞðM 21 þ M 03 Þð3ðM 30 þ M 12 Þ 2 ðm 21 þ M 03 Þ 2 Þ; I 6 ¼ðM 20 M 02 ÞððM 30 þ M 12 Þ 2 ðm 21 þ M 03 Þ 2 Þ þ 4M 11 ðm 30 þ M 12 ÞðM 21 þ M 03 Þ ð2þ I 7 ¼ M 20 M 02 M11 2 I 8 ¼ M 30 M 12 þ M 21 M 03 M12 2 M2 21 Central moments of order ðp þ qþ: M pq ¼ PN 1 PM 1 i¼0 j¼0 ði x s Þ p ðj y s Þ q f ði; jþ The co-ordinates of the region's centre of gravity ðcentroidþ: x s ¼ m10 m 00 ; y s ¼ m01 m 00 Moments of order ðp þ qþ: m pq ¼ PN 1 PM 1 i p j q f ði; jþ i¼0 j¼0 ð3þ

18 6 Z. Kuś and A. Nawrat Fig. 1 The set of ten objects which are going to be recognised

19 Minimizing the Image Resolution in Order to Increase 7 where f(i, j) image pixel, M the height of the image, N the width of the image. Ten pattern images, cf. Fig. 1, (seven images of the vehicles and three images of various terrains) define ten patterns which should be distinguished from each other. The resolution will be changed in this way that in every next step we will obtain the image of four times smaller area. It means that we will examine the cases for eight values of resolutions: original image not downscaled and image linearly downscaled in each axis: 2, 4, 8, 16, 32, 64 and 128 times. Figure 2 presents the examples of the various image resolution on the example of the selected object. Fig. 2 The examples of the various image resolution on the example of the selected objects

20 8 Z. Kuś and A. Nawrat 3 Examining the Influence of the Image Resolution on the Moment Invariants Values for Grey Scale Images To examine the influence of image resolution on pattern recognition, we will use the pattern images presented in Fig. 1. All the images will be resized for the following linear scale rates: 1/1, 1/2, 1/8, 1/16, 1/32, 1/64, 1/128. The values of the moment invariants J 1 to J 8 for all scale rates and exemplary images are presented in Figs. 3, 4, 5, 6, 7, 8, 9 and 10. Firstly, we will examine the influence of the resolution on the values of the moment invariants. Figures 3, 4, 5, 6, 7, 8, 9 and 10 present the values of J 1 to J 8 for all examined patterns and resolutions. According to Fig. 3, we can see that the values of J 1 are almost unaltered as the resolution decreases. According to Fig. 4, we can see that the values of J 2 are slightly variable for the resolutions steps which are lower than 6. When the step of resolution is higher than 5, we can observe that J 2 increases. This takes place for almost all objects. According to Fig. 5, we can see that the values of J 3 are almost constant for resolution steps lower than 5 whereas the resolution steps are variable for higher ones. According to Fig. 6, we can see that the values of J 4 are almost constant for the step of resolution lower than 7 whereas the values of J 4 change for steps of resolution equal 7 and 8. According to Fig. 7, we can see that the values of J 5 are variable when the resolution changes for all objects. Fig. 3 The graphs of the invariant function J 1 for ten objects and eight resolutions