Search

This paper presents our latest extension of the Brno Urban Dataset (BUD), the Winter Extension (WE). The dataset contains data from commonly used sensors in the automotive industry, like four RGB and single IR cameras, three 3D LiDARs, differential RTK GNSS receiver with heading estimation, the IMU and FMCW radar. Data from all sensors are precisely timestamped for future offline interpretation and data fusion. The most significant gain of the dataset is the focus on the winter conditions in snow-covered environments. Only a few public datasets deal with these kinds of conditions. We recorded the dataset during February 2021 in Brno, Czechia, when fresh snow covers the entire city and the surrounding countryside. The dataset contains situations from the city center, suburbs, highways as well as the countryside. Overall, the new extension adds three hours of real-life traffic situations from the mid-size city to the existing 10 h of original records. Additionally, we provide the precalculated YOLO neural network object detection annotations for all five cameras for the entire old data and the new ones. The dataset is suitable for developing mapping and navigation algorithms as well as the collision and object detection pipelines. The entire dataset is available as open-source under the MIT license.

In this paper, we present our new sensor fusion framework for self-driving cars and other autonomous robots. We have designed our framework as a universal and scalable platform for building up a robust 3D model of the agent's surrounding environment by fusing a wide range of various sensors into the data model that we can use as a basement for the decision making and planning algorithms. Our software currently covers the data fusion of the RGB and thermal cameras, 3D LiDARs, 3D IMU, and a GNSS positioning. The framework covers a complete pipeline from data loading, filtering, preprocessing, environment model construction, visualization, and data storage. The architecture allows the community to modify the existing setup or to extend our solution with new ideas. The entire software is fully compatible with ROS (Robotic Operation System), which allows the framework to cooperate with other ROS-based software. The source codes are fully available as an open-source under the MIT license. See https://github.com/Robotics-BUT/Atlas-Fusion. Index Terms—Open Source, Autonomous Agent, Self Driving Car, Sensor Fusion, Mapping, ROS

One of the biggest challenges of training deep neural network is the need for massive data annotation. To train the neural network for object detection, millions of annotated training images are required. However, currently, there are no large-scale thermal image datasets that could be used to train the state of the art neural networks, while voluminous RGB image datasets are available. This paper presents a method that allows to create hundreds of thousands of annotated thermal images using the RGB pre-trained object detector. A dataset created in this way can be used to train object detectors with improved performance. The main gain of this work is the novel method for fully automatic thermal image labeling. The proposed system uses the RGB camera, thermal camera, 3D LiDAR, and the pre-trained neural network that detects objects in the RGB domain. Using this setup, it is possible to run the fully automated process that annotates the thermal images and creates the automatically annotated thermal training dataset. As the result, we created a dataset containing hundreds of thousands of annotated objects. This approach allows to train deep learning models with similar performance as the common human-annotation-based methods do. This paper also proposes several improvements to fine-tune the results with minimal human intervention. Finally, the evaluation of the proposed solution shows that the method gives significantly better results than training the neural network with standard small-scale hand-annotated thermal image datasets.

Robotic Template Library (RTL) is a set of tools for dealing with geometry and point cloud processing, especially in robotic applications. The software package covers basic objects such as vectors, line segments, quaternions, rigid transformations, etc., however, its main contribution lies in the more advanced modules: The segmentation module for batch or stream clustering of point clouds, the fast vectorization module for approximation of continuous point clouds by geometric objects of higher grade and the LaTeX export module enabling automated generation of high-quality visual outputs. It is a header-only library written in C++17, uses the Eigen library as a linear algebra back-end, and is designed with high computational performance in mind. RTL can be used in all robotic tasks such as motion planning, map building, object recognition and many others, but the point cloud processing utilities are general enough to be employed in any field touching object reconstruction and computer vision applications as well.

Autonomous driving is a dynamically growing field of research, where quality and amount of experimental data is critical. Although several rich datasets are available these days, the demands of researchers and technical possibilities are evolving. Through this paper, we bring a new dataset recorded in Brno, Czech Republic. It offers data from four WUXGA cameras, two 3D LiDARs, inertial measurement unit, infrared camera and especially differential RTK GNSS receiver with centimetre accuracy which, to the best knowledge of the authors, is not available from any other public dataset so far. In addition, all the data are precisely timestamped with sub-millisecond precision to allow wider range of applications. At the time of publishing of this paper, recordings of more than 350 km of rides in varying environment are shared at: https: //github.com/RoboticsBUT/Brno-Urban-Dataset.

The RoScan is a novel, high-accuracy multispectral surface scanning system producing colored 3D models that include a thermal layer. (1) Background: at present, medicine still exhibits a lack of objective diagnostic methods. As many diseases involve thermal changes, thermography may appear to be a convenient technique for the given purpose; however, there are three limiting problems: exact localization, resolution vs. range, and impossibility of quantification. (2) Methods: the basic principles and benefits of the system are described. The procedures rely on a robotic manipulator with multiple sensors to create a multispectral 3D model. Importantly, the structure is robust, scene-independent, and features quantifiable measurement uncertainty; thus, all of the above problems of medical thermography are resolved. (3) Results: the benefits were demonstrated by several pilot case studies: medicament efficacy assessment in dermatology, objective recovery progress assessment in traumatology, applied force quantification in forensic sciences, exact localization of the cause of pain in physiotherapy, objective assessment of atopic dermatitis, and soft tissue volumetric measurements. (4) Conclusion: the RoScan addresses medical quantification, which embodies a frequent problem in several medical sectors, and can deliver new, objective information to improve the quality of healthcare and to eliminate false diagnoses.

This article presents unmanned aerial system (UAS)-based photogrammetry as an efficient method for the estimation of snow-field parameters, including snow depth, volume, and~snow-covered area. Unlike~similar studies employing UASs, this method benefits from the rapid development of compact, high-accuracy global navigation satellite system (GNSS) receivers. Our custom-built, multi-sensor system for UAS photogrammetry facilitates attaining centimeter- to decimeter-level object accuracy without deploying ground control points; this technique is generally known as direct georeferencing. The method was demonstrated at Mapa Republiky, a snow field located in the Krkonose, a mountain range in the Czech Republic. The location has attracted the interest of scientists due to its specific characteristics; multiple approaches to snow-field parameter estimation have thus been employed in that area to date. According to the results achieved within this study, the proposed method can be considered the optimum solution since it not only attains superior density and spatial object accuracy (approximately one decimeter) but also significantly reduces the data collection time and, above all, eliminates field work to markedly reduce the health risks associated with avalanches.

This article discusses the highly autonomous robotic search and localization of radiation sources in outdoor environments. The cooperation between a human operator, an unmanned aerial vehicle, and an unmanned ground vehicle is used to render the given mission highly effective, in accordance with the idea that the search for potential radiation sources should be fast, precise, and reliable. Each of the components assumes its own role in the mission; the unmanned aerial vehicle (in our case, a multirotor) is responsible for fast data acquisition to create an accurate orthophoto and terrain map of the zone of interest. Aerial imagery is georeferenced directly, using an onboard sensor system, and no ground markers are required. The unmanned aerial vehicle can also perform rough radiation measurement, if necessary. Since the map contains three-dimensional information about the environment, algorithms to compute the spatial gradient, which represents the rideability, can be designed. Based on the primary aerial map, the human operator defines the area of interest to be examined by the applied unmanned ground vehicle carrying highly sensitive gamma-radiation probe/probes. As the actual survey typically embodies the most time-consuming problem within the mission, major emphasis is put on optimizing the unmanned ground vehicle trajectory planning; however, the dual-probe (differential) approach to facilitate directional sensitivity also finds use in the given context. The unmanned ground vehicle path planning from the pre-mission position to the center of the area of interest is carried out in the automated mode, similarly to the previously mentioned steps. Although the human operator remains indispensable, most of the tasks are performed autonomously, thus substantially reducing the load on the operator to enable them to focus on other actions during the search mission. Although gamma radiation is used as the demonstrator, most of the proposed algorithms and tasks are appli

Vectorization is a widely used technique in many areas, mainly in robotics and image processing. Applications in these domains frequently require both speed (for real-time operation) and accuracy (for maximal information gain). This paper proposes an optimization for the high speed vectorization methods, which leads to nearly optimal results. The FTLS algorithm uses the total least squares method for fitting the lines into the point cloud and the presented augmentation for the refinement of the results, is based on a modified Nelder-Mead method. As shown on several experiments, this approach leads to better utilization of the information contained in the point cloud. As a result, the quality of approximation grows steadily with the number of points being vectorized, which was not achieved before. Performance costs are still comparable to the original algorithm, so the real-time operation is not endangered.

This paper presents the procedures enabling the calibration and evaluation of intrinsic parameters in a Velodyne multi-beam laser scanner. As the device will be utilized in field robotics applications, both the evaluated parameters and the calibration are generally aimed to improve the performance of the scanner with advanced mapping algorithms such as robot evidence grids or octree. The measured parameters are compared with the data provided by the manufacturer. A novel calibration method based on conditional adjustment for correlated measurements is proposed and compared with factory calibration., Tento článek prezentuje postupy umožňující kalibraci a hodnocení vnitřních parametrů 3D skeneru Velodyne. Vzhledem k tomu, že zařízení bude využito v reálných robotických aplikacích, hodnocené parametry a kalibrace jsou obecně zaměřeny na zlepšení výkonnosti skeneru s pokročilými mapovacími algoritmy. Naměřené parametry jsou porovnány s údaji poskytnutými výrobcem.Byla navržena nová kalibrační metoda založená na podmínkovém vyrovnání korelovaných měření a srovnána s tovární kalibrací.

Limb volume measurements are used for evaluating growth of muscle mass and effectivity of strength training. Beside sport sciences, it is used e.g. for detection of oedemas, lymphedemas or carcinomas or for examinations of muscle atrophy. There are several commonly used methods, but there is a lack of clear comparison, which shows their advantages and limits. The accuracy of each method is uncertainly estimated only. The aim of this paper is to determine and experimentally verify their accuracy and compare them among each other. Water Displacement Method (WD), three methods based on circumferential measures—Frustum Sign Model (FSM), Disc Model (DM), Partial Frustum Model (PFM) and two 3D scan based methods Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) were compared. Precise reference cylinders and limbs of two human subjects were measured 10 times by each method. Personal dependency of methods was also tested by measuring 10 times the same object by 3 different people. Accuracies: WD 0.3 %, FSM 2–8 % according person, DM, PFM 1–8 %, MRI 2 % (hand) or 8 % (finger), CT 0.5 % (hand) or 2 % (finger);times: FSM 1 min, CT 7 min, WD, DM, PFM 15 min, MRI 19 min; and more. WD was found as the best method for most of uses with best accuracy. The CT disposes with almost the same accuracy and allows measurements of specific regions (e.g. particular muscles), as same as MRI, which accuracy is worse though, but it is not harmful. Frustum Sign Model is usable for very fast estimation of limb volume, but with lower accuracy, Disc Model and Partial Frustum Model is useful in cases when Water Displacement cannot be used.

Existují případy, kdy jsou standardní lékařské zobrazovací metody sice schopny poskytnout dostatečné výsledky, ale ne zcela optimálním způsobem. Například se zbytečně vysokými náklady, s redundantními informacemi nebo se zbytečnými nároky na pacienta. Tento článek se zabývá návrhem nového zobrazovacího zařízení pro situace, kdy pro vyšetření potřebujeme pouze model povrchu těla, zobrazení vnitřních struktur je zbytečné. V tomto případě přináší navržené zařízení významná zlepšení optimalizace léčebného procesu. V první části je provedena analýza současně používaných lékařských zobrazovacích metod, která ukazuje, že žádné dostupné zobrazovací zařízení není pro uvedené účely vhodné. V další části je popsán vývoj nového specializovaného zobrazovacího zařízení a jsou popsány jeho principy a funkce. Parametry nového zařízení jsou porovnány se stávajícími metodami. Srovnání ukazuje, že navržené zařízení je v mnoha pohledech lepší než zařízení současně používaná., There are special medical cases, where standard medical imaging modalities are able to offer sufficient results, but not in the optimal way. It means, that desired results are produced with unnecessarily high expenses, with redundant informations or with needless demands on patient. This paper deals with one special case, where information useful for examination is the body surface only, inner sight into the body is needless. New specialized medical imaging device is developed for this situation. In the Introduction section, analysis of presently used medical imaging modalities is presented, which declares, that no available imaging device is best fitting for mentioned purposes. In the next section, development of the new specialized medical imaging device is presented, and its principles and functions are described. Then, the parameters of new device are compared with present ones. It brings significant advantages comparing to present imaging systems.

The article deals with hardware and software design of the device for human head movement simulating. Dynamics of head movement is significant, but is an attempt to achieve the same dynamic properties of the device as a human head. The article contains a complete description of the software and design of mechanical parts. Construction equipment is not complete yet., Článek se zabývá hardwarovým a softwarovým návrhem zařízení, které by co možná nejpřesněji simulovalo pohyby lidské hlavy. Dynamika pohybu hlavy je značná, přesto je snahou docílit stejných dynamických vlastností zařízení jako má lidská hlava. Článek obsahuje kompletní popis programového vybavení a návrh mechanické části. Konstrukce zařízení není dosud hotova.

Tato bakalářská práce je rozdělena do dvou částí. První část je spíše teoretická, neboť zahrnuje principy družicové navigace a metody zpřesňování polohy GPS přijímače. Druhá část je aplikační a popisuje navigační subsystém z hardwarového a softwarového hlediska. Úkolem této bakalářské práce bylo vytvořit navigační subsystém pro robota, navrhnout komunikační kanál a implementovat navigační algoritmus do robota., This bachelor’s work is devided into two parts. The first part is theoretical, because it includes the principles of satellite navigation and the metodies of giving precision to position.The second part is aplication part and it describes the navigation subsystem of hardware and software equipment. The result of this bachelor’s work was made a navigation subsystem for robot, designed communication and implemented navigation algorithm into the robot.

Cílem práce je navrhnout a realizovat řízení a programovou obsluhu létajícího robotu koncepce quadrotoru. Práce je rozdělena do tří hlavních celků. První část se zabývá studií již hotových projektů, jejich vlastnostmi a použitými komponenty. Následuje popis vlastní konstrukce a fyzikálních vlastností koncepce. Druhá část pojednává o použitých komponentách. Podrobně je rozebrán především senzorický subsystém, který se zaměřuje na obsluhu senzorů, zpracováním jejich dat pro potřeby stabilizace a eliminací chyb. Dále pak obsluha akčních členů, dálkové řízení a komunikace s pozemní stanicí. V třetí části jsou ověřeny fyzikální vlastnosti, zejména akčních členů a je navrhnut algoritmus pro základní stabilizaci. Ten se skládá z trojice nezávislých regulátorů stabilizující 3 stupně volnosti robotu, jimiž jsou rotační pohyby., The Thesis objective is the design and implementation of program management and operational service for a flying robot, quadrotor concept. This thesis is divided into three main parts. The first part is focused on the studies of finished projects, their characteristics and used components, followed by a description of structure and physical properties of the concepts. The second part is focused on the components. It mainly discusses the sensory subsystem, which is concentrated on the sensors operations, data processing for the stabilization needs and elimination of errors. Then service of actuators, remote control and the ground station communication. In the third part are verified physical properties of actuators and designed basic algorithm of stabilisation. This consists of three independent regulator, stabilizing 3 degrees of robot movement freedom, which are rotational movements.

Tato práce popisuje způsob automatického nalezení cesty ve 2D bludišti, které je zadáno formou bitmapových mřížek obsazenosti. Problém byl řešen v jazyce C# za pomoci algoritmu přímkové dekompozice, pomoci kterého je vytvořena síť potencionálních cest. Tento algoritmus předpokládá aproximaci překážek vektorovými polygony. Pro nalezení výsledné cesty je použit algoritmus A*. Výstupem je cesta popsaná úsečkami, kterou by po dalších úpravách bylo možné použít pro navigaci robota., This thesis describes an automatic path finding method in a 2D maze defined by bitmap evidence grids. The problem was solved in the programming language C# with the help of the straight line decomposition algorithm used for creating a network of potential routes. The algorithm assumes approximation of obstacles by vector polygons. The resulting path was found by the means of algorithm A*. The result is a path described by line segments. After some modifications, the path could be used for robot navigation.

Práca je rozdelená do niekožkých logických celkov. V prvom je prebratá problematika mobilnej robotiky, konštrukcia, pohony a riadenie mobilných robotov. Ďalšie časti sa venujú postupne návrhom mechanickej konštrukcie robota, výpočtom požiadaviek na motory a návrhom typu pohonov. Nakoniec je navrhnutá riadiaca elektronika pre tieto motory spolu s programom pre mikropočítač., The thesis is divided into several logical units. In the first part is described the topic of mobile robotics, construction, propulsion and control of mobile robots. Other parts are devoted to the design of robot’s mechanical structure, calculating the requirements for the proposed type of engines and drives. Finally, the control electronics for these engines is proposed, along with a program for microcomputer.

Cílem této práce je prozkoumat snímače náklonu používané v mobilních zařízení Android a porovnat je s přesným referenčním senzorem Xsens. Dalším z cílů bylo vytvořit Android aplikaci na měření náklonu balanční plošiny, která komunikuje přes Bluetooth s počítačem. Z měření vyplývá, že snímače v mobilních telefonech jsou dostatečně kvalitní, aby mohli nahradit přesný referenční senzor Xsens pro měření náklonů balanční plošiny. Na základě těchto výsledků byla vytvořena aplikace. Mobilním telefonem s touto aplikací lze senzor Xsens nahradit, čímž dojde k výraznému zlevnění celého řešení, a tudíž možnému rozšíření mezi pacienty a rehabilitační zařízení., The aim of this thesis is to evaluate sensors used in Android mobile phones for measuring inclination and their comparation with precise reference sensor Xsens. Another goal is to create Android application for measuring inclination of balance desk. This application communicates with a computer by Bluetooth. Results indicates that sensors in mobile phones are sufficient enough to replace the precise sensor Xsens for measuring inclination of balance desk. Based on the results mobile application was created. Using a mobile phone with this application can replace Xsens, which results in rapid decrease of the price and the possible use in rehabilitation centers or by patients.

Tato bakalářská práce je rozdělena do dvou částí. První část je spíše teoretická, neboť zahrnuje principy družicové navigace a metody zpřesňování polohy GPS přijímače. Druhá část je aplikační a popisuje navigační subsystém z hardwarového a softwarového hlediska. Úkolem této bakalářské práce bylo vytvořit navigační subsystém pro robota, navrhnout komunikační kanál a implementovat navigační algoritmus do robota., This bachelor’s work is devided into two parts. The first part is theoretical, because it includes the principles of satellite navigation and the metodies of giving precision to position.The second part is aplication part and it describes the navigation subsystem of hardware and software equipment. The result of this bachelor’s work was made a navigation subsystem for robot, designed communication and implemented navigation algorithm into the robot.

Tato bakalářská práce se zabývá návrhem řídicí a komunikační jednotky pro létající robot známý pod názvem Quadrocopter. Cílem práce je návrh a praktická realizace jednotky, která umoţní měření náklonu a natočení robota ve vzdušném prostoru pomocí inerciálního snímače. Dále pak tato jednotka zprostředkuje uţivateli informace o letové výšce, v jaké se robot nachází. Všechna naměřená data jsou přenášena přes komunikační moduly na zem a zpracovávána uţivatelským softwarem., This thesis deals with design control and communication unit for flying robot known as Quadrocopter. The goal is design and practical implementation unit, which allows measurement of tilt and rotation of the robot in the airspace using inertial sensors. Furthermore, this unit conveys the user information about the flight altitude at which the robot is located. All measured data are transmitted via communication modules to the ground and processed user software.

Tato práce se zabývá dosavadním vývojem pohonů v robotice, jejich aplikacemi, specifiky různých pohonů jejich principy, výhodami a nevýhodami. Práce je rozdělena do části věnující se běžně využívaným pohonům v robotice, jako jsou motory na elektrickém principu, v dnešní době zejména PMSM motory, dále pohony na pneumatickém či hydraulickém principu využívající k převodu tlakové energie pístového principu, také motory pístové spalovací. V následující části se zaobírá méně tradičními pohony, které mají velmi zajímavé vlastnosti pro použití. Většina těchto pohonů je přirozeně pružná a deformovatelná a mají mnohem lepší poměr výkonu ku hmotnosti. V poslední části je vysvětlen princip funkce pneumatického krokového motoru, který je netradičním pohonem použitelným pro aplikaci v robotu pracujícím pod magnetorezonančním zobrazováním., This report presents up to now development of actuators in robotics theirs applications, specifics various sorts of actuators, theirs principles, advantages and disadvantages. Report is splitted to the part that presents usual used actuators, especially electric principle powered actuators, today primarily PMSM motors, furthermore pneumatic or hydraulic actuators that uses pistons principle for convert pressure energy to the motion, or piston combustion engines. In next part presents more unusual actuators that have interesting features for use. Lot of these actuators are elastic, deformable and have better power/weight ratio. In the last part is presented priciple of function pneumatic stepper motor, that is uncommon actuator useable for application in robot works compatible with magnetic resonance imaging.