Large Number Of Objects Research Articles

AI IN AUTONOMOUS VEHICLES: SAFETY AND CHALLENGES

Artificial intelligence in autonomous vehicles aims to ensure the safety of vehicles on public roads. An autonomous vehicle (AV) coordinates complex insight and limitation parts to make a model of its general surroundings, which is then used to explore the vehicle securely. Ai (ml) based models are unavoidably utilized in these parts to extricate objects in- arrangement from uproarious sensor information. The prerequisites for these parts are essentially set to accomplish as high as could really be expected. With current avs conveying numerous sensors (cameras, radars, and lidar), handling every one of the pieces of information in constant prompts engineers making compromises which may bring about a less than ideal framework in specific driving circumstances. Because of the absence of exact necessities on individual components, secluded testing and approval additionally becomes challenging. In this paper, we plan the issue of determining stomach muscle extract world model precision required for safe av conduct from high level driving situation recreations. This is computationally costly as the world model can contain a large number of objects with a few credits and an av extricates a world model each time-step during the reproduction. We depict ways to deal with effectively address the issue and infer part level necessities and tests.

The Natural Bases of Neutrosophy

Neutrosophy began as a branch of philosophy that considered neutrality in addition to the positive and negative. It consists of the addition consideration of a neutral state to complement the binary approach of true or false. Its creator quickly extended it to the field of mathematics and it was gradually applied to all fields of science. Here, we present a reverse approach that highlights the importance of neutrality in all fields of study and application, citing some revealing examples. Furthermore, we explain that this importance of neutrality is intrinsic to all sciences because it is based on natural foundations. Indeed, neutrality is a forming part first of all of the human conception of things, of our way of thinking, of cognition in general but also of living things, matter and even particles. In addition to these most real-world physical concrete aspects, neutrality is inherent to mathematics, to logic first of all, but also to probabilities and statistics where neutrality which simply results from a large number of objects, the universe. Thus neutrosophy is well adapted to the majority of applied problems because its modeling is inspired by reality and that it allows, in particular, to deal with the component of uncertainty and indeterminacy that the real world comprises intrinsically.

Monitoring of vegetation cover of rewetted peatlands in Moscow Oblast

Abandoned drained peatlands pose the greatest danger in terms of peat fires. This is especially true of abandoned, unrestored peat milling fields, the maximum areas of which are located in Moscow Oblast. The most effective way to reduce the fire hazard and improve the ecological condition of such lands in the absence of economic reasons for their use is rewetting and peatland restoration. It is necessary to prioritize objects for rewetting and to monitor the effectiveness of measures taken. Such objects can be of considerable size, difficult to pass and not amenable to ground mapping. Fires and rewetting contribute to rapid changes in vegetation cover. Their monitoring requires remote sensing data and, above all, multispectral satellite data. The large number of objects, their spatial distribution and the need for multi-year observations require a combination of data from different satellites. We present the experience of monitoring more than 73 thousand hectares of peatlands in Moscow Oblast rewetted in 2010-2013 after the fires of 2010. Current results show a reduction in the areas of open peat and dry grass vegetation, insignificant changes in the areas occupied by coniferous trees and progressive expansion of the area occupied by deciduous trees and shrubs, as well as an increase of hydrophilic communities and open water after rewetting and subsequent maintenance of the area. The developed and tested methodology makes it possible to assess the effectiveness of peatland rewetting to reduce the risk of peat fires, as well as to identify areas requiring additional measures. The proposed algorithm ensures reliable allocation in semi-automatic mode of 6 classes of vegetation/land cover, different in fire hazard and basic environmental characteristics that determine the fluxes of greenhouse gases, biodiversity and other ecosystem functions.

Spatial-Temporal Assessment of Formation of Nature Reserve Fund of Sumy Region

Purpose. To assess the main trends in the formation of nature reserves and the degree of protection of the territory of Sumy region to determine the directions of development of the regional ecological network. Methods. Methods of historical analysis, statistical processing, spatial modeling, evaluation and generalization of the obtained data are used. Results. During the study period since 1991, the area of ​​NPF has increased 5.44 times, and the number of NPF objects - almost 2 times and in 2019 is 275 units. Comparison of administrative districts of Sumy region by the number of NPF objects allows to determine the districts - "leaders": Romensky (37), Sumy and Lebedyn (25 pieces each), Krolevets (24) and districts - outsiders - Buryn (5), Shostka (7) and Velykopysarivsky 8). By the area of ​​the nature reserve fund "leaders" - Putivl district (39298.55 hectares) and Konotop district (32440.7 hectares). The discrepancy between a large number of objects and the area of ​​the NPF prompted the determination of the percentage of reserves and modeling of the relationship between these three indicators. Based on cartographic visualization and statistical modeling, recommendations have been developed, the main of which are: intensification of the creation of NPF objects of local significance, first of all, in the areas that are most "lagging behind" in the process of bequest; renewal of previously rejected NFP Projects and facilitate their approval; initiating activities to promote tourist routes on unique natural, historical, ethnographic sites of the NPF for economic support of communities. Conclusions. It is established that the degree of protection of the territory of Sumy region is lower than recommended and averages 7.43%. Spatially, there is a band of high concentration of NPF objects with relatively large areas, which even exceeds the recommended (Konotop, Krolevets and Putivl, or according to the new zoning - Krolevets district) in the central part of the region, which is surrounded on both sides by areas with low percentage (up to 5%) of protected areas and only the extreme northern and southern parts of the region (Seredino-Budsky, Trostyanetsky, Velykopysarivsky and Okhtyrsky districts) are approaching the norm in terms of the percentage of protected areas. The patterns identified by modeling allow us to recommend increasing the share of NPF in the total area of ​​the region in two directions - increasing the area of ​​the existing facilities or creating a large number of new ones with small areas.

Efficient Method to find Nearest Neighbours in Flocking Behaviours

Flocking is a behaviour in which objects move or work together as a group. This behaviour is very common in nature think of a flock of flying geese or a school of fish in the sea. Flocking behaviours have been simulated in different areas such as computer animation, graphics and games. However, the simulation of the flocking behaviours of large number of objects in real time is computationally intensive task. This intensity is due to the n-squared complexity of the nearest neighbour (NN) algorithm used to separate objects, where n is the number of objects. This paper proposes an efficient NN method based on the partial distance approach to enhance the performance of the flocking algorithm and its application to flocking behaviour. The proposed method was implemented and the experimental results showed that the proposed method outperformed conventional NN methods when applied to flocking fish.

ON THE STUDIES OF THE ZOOMORPHIC IMAGES REPERTOIRE IN THE ART OF INKARDARIA SAKAS

The article is devoted to zoomorphic images analysis received from the Saka sites of the Eastern Aral Sea region — Uygarak (VII—VI cc. BC) and South Tagisken (VII—V cc. BC). The images were studied by the Khorezm archaeological and ethnographic expedition under the guidance of S. P. Tolstov. This material has been due to the fact that now there is an update of the source base of the studies of art objects belonging to the Saka time in the archaeology of the early Iron Age of Kazakhstan.
 At present the materials on Uygarak and South Tagisken are still among the sites with a large series of excavated mounds, both in the region and in Kazakhstan. Almost completely excavated burial grounds of the early Iron Age containing a large number of objects in Kazakhstan archaeology are single.
 The analysis of the material on Uygarak and Tagisken depicts that objects with zoomorphic decor were found in every burial. In percentage this makes 21 % in Uygarak and 18 % in Tagisken. According to the archaeologists O. A. Vishnevskaya, M. A. Itina, L. T. Yablonsky, who studied the monuments, the mounds were robbed by contemporaries. Consequently, conducting an objective analysis is difficult, but a trial to trace the dynamics of the emergence of objects created in zoomorphic style seems extremely interesting. There is a high probability that a small number of objects with preserved works of ancient art marks complexes belonging to the steppe aristocracy.
 Among zoomorphic images, the most common were birds and cat predators. There are objects in the shape of saigas and horses. Images of bears, wolves, deer, argali, wild boars, camels are quite rare. It is worth noting that bird images have never been found in Tagisken materials, and not a single horse and saiga image has been recorded in Uygarak. At the same time, the majority of zoomorphically decorated items are parts of horse equipment.

Methods for the determination of effective management decisions in insufficient information conditions

The article analyzes the methods of removing uncertainty in conditions of insufficient information for making effective management decisions in complex organizational systems. In such tasks, the choice of an effective solution depends on the state of the external environment of the study and implies the use of a number of methods or mathematical models for solving multi-criteria problems. The main methods used in cases of occurrence of a situation of uncertainty due to insufficient information are the a priori ranking methods and multiple regression analysis. However, methods based on expert judgment have a significant degree of subjectivism, and methods formed on the experimental basis and mathematical planning of research are laborious and generate uncertainty with a large number of objects or elements involved in the situation. An alternative to the above is the method of zoning, which implies dividing the set of possible states of the research environment (nature) into subsets of the dominance of individual actions. The zoning operation is an inverse parametric linear programming problem. Consequently, for solving problems in conditions of insufficient information, one can use vector optimization methods.

MULTIPLE VEHICLES AND PEOPLE TRACKING IN AERIAL IMAGERY USING STACK OF MICRO SINGLE-OBJECT-TRACKING CNNS

Abstract. Geo-referenced real-time vehicle and person tracking in aerial imagery has a variety of applications such as traffic and large-scale event monitoring, disaster management, and also for input into predictive traffic and crowd models. However, object tracking in aerial imagery is still an unsolved challenging problem due to the tiny size of the objects as well as different scales and the limited temporal resolution of geo-referenced datasets. In this work, we propose a new approach based on Convolutional Neural Networks (CNNs) to track multiple vehicles and people in aerial image sequences. As the large number of objects in aerial images can exponentially increase the processing demands in multiple object tracking scenarios, the proposed approach utilizes the stack of micro CNNs, where each micro CNN is responsible for a single-object tracking task. We call our approach Stack of Micro-Single- Object-Tracking CNNs (SMSOT-CNN). More precisely, using a two-stream CNN, we extract a set of features from two consecutive frames for each object, with the given location of the object in the previous frame. Then, we assign each MSOT-CNN the extracted features of each object to predict the object location in the current frame. We train and validate the proposed approach on the vehicle and person sets of the KIT AIS dataset of object tracking in aerial image sequences. Results indicate the accurate and time-efficient tracking of multiple vehicles and people by the proposed approach.

Cross-scale fusion detection with global attribute for dense captioning

As a new task of image understanding, the dense caption model needs to locate and describe a salient region in the image simultaneously. It inevitably divides the dense caption model into two parts, one part for detecting the regions of interest and the other part for generating regional language caption. Previous methods are relatively simple to deal with these two parts, using the feature map on the last convolution layer of RPN network to predict object coordinates, and using LSTM for regional language modeling. However, the structure of RPN is insufficient to deal with a large number of objects in the complex dataset, and LSTM also fails to effectively utilize the global information of images in regional language training, which brings opportunity to improve the performance of dense caption. In this paper, we propose a novel Cross-scale Fusion with Global Attribute model (CSGA) that enables the two parts of the dense caption model to perform normal end-to-end training without mutual interference. Furthermore, our model uses a one-stage object detector with feature map fusion operation across multiple detection scales to improve the quality of object detection part, and combines image features with the global high-level attribute to improve regional language training. We design a variety of model architectures and conducted sufficient experiments. Empirical results on Visual Genome dataset show that our model achieves competitive results with mAP 8.33.

Large-Scale Visual Relationship Understanding

Large scale visual understanding is challenging, as it requires a model to handle the widely-spread and imbalanced distribution of 〈subject, relation, object〉 triples. In real-world scenarios with large numbers of objects and relations, some are seen very commonly while others are barely seen. We develop a new relationship detection model that embeds objects and relations into two vector spaces where both discriminative capability and semantic affinity are preserved. We learn a visual and a semantic module that map features from the two modalities into a shared space, where matched pairs of features have to discriminate against those unmatched, but also maintain close distances to semantically similar ones. Benefiting from that, our model can achieve superior performance even when the visual entity categories scale up to more than 80,000, with extremely skewed class distribution. We demonstrate the efficacy of our model on a large and imbalanced benchmark based of Visual Genome that comprises 53,000+ objects and 29,000+ relations, a scale at which no previous work has been evaluated at. We show superiority of our model over competitive baselines on the original Visual Genome dataset with 80,000+ categories. We also show state-of-the-art performance on the VRD dataset and the scene graph dataset which is a subset of Visual Genome with 200 categories.

Методы визуализации протяженных ландшафтов в тренажерно-обучающих системах

One of the tasks that arise during the development of training (simulation) systems is the creation of a subsystem for displaying the earth's surface. Subsystems of this class provide modeling and visualization of the landscape, underlying surface (rivers, roads, forests, etc.), geometric objects that simulate cities or airports and so on. There are two main approaches to the visualization of extended landscapes --- continuous level of detail algorithms that use clustered triangulation and geometry clipmapbased algorithms. The method of generation and visualization of the Earth's surface presented in the work is based on the second approach. It is possible to perform calculations on the GPU using single precision, which allows faster calculations compared to using data types with double precision. In addition, this approach may be used on mobile graphics processors that do not support double precision. The proposed new method of dynamic resource management reduces the occupied video memory, which allows more detailed texture data to be loaded for a larger number of objects simultaneously.

Constrained trajectory simplification with speed preservation

ABSTRACTThe rapid increase in movement trajectory data causes data storage, transmission, computational processing, and visualization problems. These issues can be alleviated through trajectory simplification, which removes unnecessary details from raw trajectories. Most existing studies that focused on trajectory simplification considered freely moving objects, and they attempted to minimize position errors in their simplified representations in a two-dimensional plane. However, a large number of objects move within the constraint of road networks. In such constrained trajectory simplification, position error should be measured in the network space. Moreover, constrained trajectories contain a wealth of speed-change information that reflects the movement patterns of moving objects. In this study, we designed a data model, proposed error measurements, and developed a two-component method to simplify constrained trajectories. The geometric component in our method extended the classic Douglas–Peucker method using network distance to simplify trajectories with a guaranteed position error bound in network space. The semantic component enhanced the simplified representation by employing a data-enrichment strategy that allows users to control speed loss. Real trajectory data were used to assess the effectiveness of the proposed method. Experimental results show that our method has a lower position error than existing algorithms do when road network constrained trajectory data are simplified. The method can also preserve original speed in the simplified representation with a relatively low increase in data size. Our study thus provides an approach to simplifying trajectory data that guarantees error bounds in both location and speed.

Completeness: How the Lack of a Mouse in a Box Revisits the Spectacle of the Kunstkammer

Set aside your expectations that this will be your typical museum visit – instead, prepare for a disruptive and engaging experience (Figure 1). ‘Spitzmaus Mummy in a Coffin and Other Treasures’ at the Kunsthistorisches Museum (KHM) in Vienna and curated by Wes Anderson and Juman Malouf together with Jasper Sharp and Mario Mainetti is an exhibition that leaves visitors either perplexed or enchanted.

Fusing incomplete preference rankings in design for manufacturing applications through the ZM II -technique

The authors recently presented a technique (denominated “ZM”) to fuse multiple (subjective) preference rankings of some objects of interest—in manufacturing applications—into a common unidimensional ratio scale Franceschini and Maisano J Intell Manuf, 2019. Although this technique can be applied to a variety of decision-making problems in the manufacturing field, it is limited by a response mode requiring the formulation of complete preference rankings, i.e., rankings that include all objects. Unfortunately, this model is unsuitable for some practical contexts—such as decision-making problems characterized by a relatively large number of objects and field surveys—where respondents can barely identify the more/less preferred objects, without realistically being able to construct complete preference rankings. The purpose of this paper is to develop a new technique (denominated “ZMII”) which also “tolerates” incomplete preference rankings, e.g., rankings with the more/less preferred objects only. This technique borrows the underlying postulates from the Thurstone’s Law of Comparative Judgment and uses the Generalized Least Squares method to obtain a ratio scaling of the objects of interest, with a relevant uncertainty estimation. Preliminary results show the effectiveness of the new technique even for relatively incomplete preference rankings. Description is supported by an application example concerning the design of a coach-bus seat.

Many-objective differential evolution optimization based on reference points: NSDE-R

Design methodologies of today require the solution of several many-objective optimization problems. The last two decades have seen a surge in several algorithms capable of solving multi-objective optimization problems. It was only in the past 5 years that new algorithms capable of coping with a large number of objectives have been introduced. This work presents a new differential evolution algorithm (NSDE-R) capable of efficiently solving many-objective optimization problems. The algorithms make use of reference points evenly distributed through the objective function space to preserve diversity and aid in multi-criteria-decision-making. The proposed NSDE-R was applied to test problems from the DTLZ and WFG suite, having three to 15 objectives. Two mutation donor operators were investigated for their ability to converge to the analytical Pareto front while maintaining diversity. The ability of NSDE-R to converge to a user-specified region of the Pareto front is also investigated. The proposed NSDE-R algorithm has shown to have a higher rate of convergence and better convergence to the analytical Pareto front.

Computationally efficient approach for identification of fuzzy dynamic groundwater sampling network.

In the present work, uncertainty-based dynamic sampling frameworks were tested for a sampling horizon with multiple time steps. Future concentration values were assumed as fuzzy numbers. Multiple realization-based simulations were used for generation of fuzzy numbers. The first framework considers fuzzy variance reduction. The second framework considers mass estimation error reduction with maximization of spatial coverage of the dynamic sampling network. It is a multi-objective optimization problem with a large number of objectives. In Dhar and Patil (2012), uncertainty-based optimal sampling design model was suggested using Nondominated Sorting Genetic Algorithm-II (NSGA-II) as its optimization algorithm. However, NSGA-II becomes computationally expensive while handling more than three objectives. We extend the previously suggested algorithm for multi-objective sampling network design problems based on NSGA-III framework. Two design frameworks were proposed: one incorporating a simulation model and a fuzzy covariance for minimizing the total contaminant-concentration variance and the other incorporating a simulation model and a fuzzy kriging model in conjunction with an optimization model to minimize the fuzzy mass estimation error and spatial coverage of spatiotemporal sampling locations. NSGA-III was used for solving the sampling network design model. Performances of the proposed frameworks were evaluated for two hypothetical illustrative examples. The results indicate that the proposed design frameworks perform satisfactorily under uncertain system conditions.

Extent Pursuit for Cross-Sectional Dependence in Large Panels

Accurate estimation for extent of cross-sectional dependence in large panel data analysis is paramount to further statistical analysis on the data under study. Grouping more data with weak relations (cross--sectional dependence) together often results in less efficient dimension reduction and worse forecasting. This paper describes cross-sectional dependence among a large number of objects (time series) via a factor model and parametrizes its extent in terms of strength of factor loadings. A new joint estimation method is proposed for the parameter representing such extent and some other parameters involved in the estimation procedure. In particular, a marginal estimator is also proposed as the related other parameters are observed. Asymptotic distributions for the joint estimator and marginal estimator are both established. Various simulation designs illustrate the effectiveness of the proposed estimation method in the finite sample performance. Applications in cross-country macro-variables and stock returns from S&P 500 are studied.

Task planning in robotics: an empirical comparison of PDDL- and ASP-based systems

Robots need task planning algorithms to sequence actions toward accomplishing goals that are impossible through individual actions. Off-the-shelf task planners can be used by intelligent robotics practitioners to solve a variety of planning problems. However, many different planners exist, each with different strengths and weaknesses, and there are no general rules for which planner would be best to apply to a given problem. In this study, we empirically compare the performance of state-of-the-art planners that use either the planning domain description language (PDDL) or answer set programming (ASP) as the underlying action language. PDDL is designed for task planning, and PDDL-based planners are widely used for a variety of planning problems. ASP is designed for knowledge-intensive reasoning, but can also be used to solve task planning problems. Given domain encodings that are as similar as possible, we find that PDDL-based planners perform better on problems with longer solutions, and ASP-based planners are better on tasks with a large number of objects or tasks in which complex reasoning is required to reason about action preconditions and effects. The resulting analysis can inform selection among general-purpose planning systems for particular robot task planning domains.

Design and Implementation of a Mixed IoT LPWAN Network Architecture.

IoT is much more than a large number of objects or customer devices connected to the Internet. IoT offers organizations many more opportunities than they can imagine. According to this, sooner or later they will probably choose to build their own IoT network. In this article, we review the technologies of IoT LPWAN Sigfox and LoRa. It can be considered the most important at present due to its ability to make the smart city possible. We also propose the development, deployment and implementation of a mixed IoT architecture LoRa-Sigfox composed of components based on open hardware and software. The architecture is evaluated in a real environment focused on remote monitoring of water meter devices.

TRANSCENDENTAL MATERIALISM OF METHOD

The classical concept of method is based on the cognitive ability of the “object of the living” and serves as a rationale for the relationship between the phenomena of sensation and thinking in the processes of interaction of the thinking being with an indefinitely large number of objects in a given area of existence. At the same time, knowledge of the signs of the existence of any object of concrete existence, which has a complex structure of regular signs of changes in the states and properties of its existence, is difficult in knowledge. In this regard, the subject of “pure abstraction” should be considered as the simplest in knowledge, knowledge of the signs of changing states and properties of which cannot be represented in the experience of sensations and thinking of the “object of the living”. In addition, in the worldview of a specific existence in a given area, the concept of a method is defined as a path from ignorance to knowledge, from simple knowledge to complex knowledge. And as a way to overcome this path is the “ascent” from the “abstract knowledge” to the “knowledge of the concrete”. However, the accepted terminology in the definition of the method shows a sign of contradiction due to the fact that in the concept of “abstract knowledge” there can be no signs of changes in the states and properties of the object, which could be felt and conceivable in the experience of the “living thing”. Because “abstract knowledge” by its definition can only be an attribute of the subject of “pure abstraction”, knowledge of the signs of changing states and properties of which are impossible in the experience of sensations and thinking of the thinking being. Moreover, the term “abstract”, which determines the knowledge of the “object of the simple” in the experience of thinking beings, can only be a term that defines the ratio of the ability of the representations of the being thinking about the object of “simple knowledge”, showing signs of concrete existence in a given area. Therefore, in the representation of transcendental materialism in the concept of the method is the possibility of “mental reincarnation” of complex knowledge about the signs of changes in the states and properties of the object of concrete existence, by “descending” to its “sources” in the signs of changes in the states and properties of the object “pure abstraction”. From the “knowledge of the complex” to the “knowledge of the simple”.