Informatics and Automation

The study introduces an innovative intelligent agent-controlled elevator system specially designed to improve passenger wait times and enhance the efficiency of high-rise buildings. By utilizing the classic single-agent planning model, we developed a unique strategy for handling calls from halls and cars, and combined with this strategy we significantly improved the overall performance of the elevator system. Our intelligent control methods are in-depth compared with conventional elevator systems, assessing three important performance indicators: response time, system capacity to handle multiple active elevator cars simultaneously, and average passenger waiting time. The results of the full simulation show that an intelligent agent-based model consistently exceeds conventional elevator systems in all measured criteria. Intelligent control systems have significantly reduced response times, and improved simultaneous elevator management and passenger wait times, especially during high traffic. These advances not only improved traffic flow efficiency, but also greatly contributed to passenger satisfaction and brought smoother and more reliable transport experiences within the building. Furthermore, the increased efficiency of our systems is in line with the goals of building energy management, as it minimizes unnecessary movements and idle time. The results demonstrate the system's ability to meet dynamic, high-occupation environment requirements and mark a significant step forward in intelligent infrastructure management.

In coordinated circular motion of a group of autonomous unmanned aerial vehicles (UAVs or drones), it is important to ensure that collisions between them are avoided. A typical situation occurs when one of the drones in a circular formation needs to overtake the drone ahead. The reason for such an overtake may be due to a given geometry of the UAV formation, when this configuration of a given relative position of the drones has changed for some reason. In this case, the limited maneuverability of UAVs of exactly fixed-wing type requires taking into account the peculiarities of their dynamics in the synthesis of the collision avoidance algorithm. The impossibility of the airspeed for a fixed-wing type UAV to drop below a certain minimum value also plays a role here. In this paper, we propose to use an approach based on vortex vector fields, which are essentially a rotational modification of the artificial potential field (APF) method. In this case, the path following algorithm developed in our previous works provides the circular motion. As a result, a collision avoidance algorithm has been developed that works efficiently by maintaining a coordinated circular motion of the autonomous drone formation without unnecessary turns. The proposed algorithm was named Artificial Potential Field for Circular Motion (abbreviated as APFfCM). Using the direct Lyapunov method, it is shown that the trajectories of the formation system have uniform boundedness (UB) when using the proposed control algorithm. Due to the boundedness of the candidate Lyapunov function, it is guaranteed that no collision event between drones will occur. Thus the control objective of providing coordinated circular motion for an autonomous fixed-wing type drone formation without collisions is achieved. Fixed-wing (“flying wing”) UAV models in MATLAB/Simulink environment demonstrate the effective performance of the proposed algorithm. These models have both full nonlinear dynamics and implementation of tuned autopilots stabilizing angular and trajectory motion.

Currently, there is a significant increase in information security incidents related to attacks on web resources. Obtaining unauthorized access to web resources remains one of the main methods of penetration into corporate networks of organizations and expanding the capabilities of intruders. In this regard, many studies are aimed at developing web backdoor detection systems (WBDS), but there is a need to assess the effectiveness of these systems. The purpose of this study is to develop an objective approach to assess the effectiveness of the WBDS functioning. In this work, it was found that the effectiveness of web backdoor detection systems is objectively manifested in the process of their use, therefore, testing of such systems should be carried out in conditions as close as possible to real ones. In this regard, the article proposes a new technique for assessing the effectiveness of WBDS. It is based on the calculation of three groups of specific indicators characterizing the potency, resource intensity and responsiveness of the detection tool, as well as the calculation of a generalized effectiveness indicator. Based on an analysis of research in this area, a classification of web backdoors embedded by an attacker into the source code of web applications has been developed. This classification is used when generating test datasets to calculate specific potency indicators. The developed methodology is applicable to tools that work based on the analysis of the source code of web pages. Additionally, its use requires a number of initial data, such as permissible maximum errors of frequent potency indicators and the probability of them being within the confidence interval, as well as weighting coefficients of specific potency indicators, which are selected by expert methods. This work may be useful for information security specialists and researchers who want to conduct a more objective assessment of their WBDS.

Most research in the field of voice presentation attack detection relies on the speaker-independent approach. Nevertheless, several scientific works indicate that using the speaker-specific approach, which involves utilizing prior knowledge about the identity of the claimed speaker to enhance the accuracy of spoofing detection, is likely to be beneficial. Therefore, the goal of this work is to propose a speaker-specific method of spoofing attack detection based on anomaly detection and to evaluate its applicability to the detection of synthesized speech and converted voice. Artificial neural networks pre-trained for the tasks of spoofing detection, speaker recognition, and audio pattern recognition are used for feature extraction. A set of anomaly detection models are used as backend classifiers. Each of them is trained on bonafide data of a target speaker. The experimental evaluation of the proposed method on the ASVspoof 2019 LA dataset shows that the best speaker-specific spoofing detection system, which uses an anomaly detection model and a neural network pre-trained for the task of speaker recognition, achieves an EER of 4.74%. This result suggests that embeddings extracted by networks pre-trained for speaker recognition contain information that can be utilized for spoofing detection. In addition, the proposed method allowed to increase the accuracy of three baseline systems pre-trained for the task of spoofing detection. Experiments with two baseline systems on the ASVspoof 2019 LA dataset showed relative improvement in terms of EER by 7.1% and 9.2%, and in terms of min t-DCF by 4.6%. Experiments with the third baseline system on the ASVspoof 2021 LA dataset showed relative improvement in terms of EER by 3.9% without significant improvement of min t-DCF.

The concept of automation of the process of annotation of scientific materials (Russian-language scientific articles) is proposed and its practical implementation is carried out by means of machine learning technologies, and additional training of large language models. The relevance of correct and rational compilation of annotations is indicated, and the problems related to establishing a balance between the time-consuming process of annotation and ensuring compliance with key requirements for annotation are highlighted. The basics of annotation presented in the family of standards on information, librarianship, and publishing are analyzed, and the classification of annotations and requirements for their content and functionality is given. The essence and content of the annotation process, and the typical structure of the research object are presented schematically. The issue of integration of digital technologies into the annotation process is analyzed, and special attention is paid to the advantages of introducing machine learning and artificial intelligence technology. The digital toolkit used to generate text in natural language processing applications is briefly described. Its shortcomings for solving the problem posed in this scientific article are noted. The research part substantiates the choice of the machine learning model used to solve the problem of conditional text generation. The existing pre-trained large language models are analyzed and, considering the problem statement and existing limitations of computing resources, the ruT5-base model is selected. A description of the dataset is given, including scientific articles from journals included in the list of peer-reviewed scientific publications in which the main scientific results of dissertations for the degrees of candidate and doctor of science should be published. The data labeling technique based on the operation of the tokenizer of the pre-trained large language model is characterized, and the numerical characteristics of the dataset distributions and the parameters of the training pipeline are presented graphically and in tables. The ROUGE quality metric is used to evaluate the model, and the expert assessment method, including grammar and logic as basic criteria, is used to evaluate the results. The quality of automatic annotation generation is comparable to real texts and meets the requirements of information content, structure and compactness. The article may be of interest to an audience of scientists and researchers seeking to optimize their scientific activities in terms of integrating digitalization tools into the process of writing articles, as well as to specialists involved in training large language models.

In this study, we present a novel model called ADA-NAF (Anomaly Detection Autoencoder with the Neural Attention Forest) for semi-supervised anomaly detection that uniquely integrates the Neural Attention Forest (NAF) architecture which has been developed to combine a random forest classifier with a neural network computing attention weights to aggregate decision tree predictions. The key idea behind ADA-NAF is the incorporation of NAF into an autoencoder structure, where it implements functions of a compressor as well as a reconstructor of input vectors. Our approach introduces several technical advances. First, a proposed end-to-end training methodology over normal data minimizes the reconstruction errors while learning and optimizing neural attention weights to focus on hidden features. Second, a novel encoding mechanism leverages NAF’s hierarchical structure to capture complex data patterns. Third, an adaptive anomaly scoring framework combines the reconstruction errors with the attention-based feature importance. Through extensive experimentation across diverse datasets, ADA-NAF demonstrates superior performance compared to state-of-the-art methods. The model shows particular strength in handling high-dimensional data and capturing subtle anomalies that traditional methods often do not detect. Our results validate the ADA-NAF’s effectiveness and versatility as a robust solution for real-world anomaly detection challenges with promising applications in cybersecurity, industrial monitoring, and healthcare diagnostics. This work advances the field by introducing a novel architecture that combines the interpretability of attention mechanisms with the powerful feature learning capabilities of autoencoders.