ISSN 0236-235X (P)
ISSN 2311-2735 (E)

Journal influence

Higher Attestation Commission (VAK) - К1 quartile
Russian Science Citation Index (RSCI)

Bookmark

Next issue

4
Publication date:
09 September 2024

Journal articles №4 2015

11. The problem of creating high temperature computing systems [№4 за 2015 год]
Authors: Bobkov S.G. (bobkov@cs.niisi.ras.ru) - Federal State Institution "Scientific Research Institute for System Analysis of the Russian Academy of Sciences" (SRISA RAS) (Professor, Deputy Director);
Abstract: The article considers the problems of creating high-temperature electronic components for oil, gas, aviation and space industries. It also describes the ways to develop Intelligent Wells (IW) systems. Creation of domestic Intelligent Wells systems would solve not only an import substitution problem; it would also allow reducing the equipment cost significantly. Thus, the cost of electronically controlled valves near perforations of an Intelligent Wells could be reduced from several hundreds of thousands dollars to several tens of thousands. The space industry has similar problems. The price of Western space-related computer can reach half a million euros, and our domestic analogue can be 10 times cheaper. In addition to high radiation resistance, flights to deep space will require functioning in a wide temperature range. The paper considers basic SRISA chips for space applications with functioning temperature up to +125°C, as well as the ways to increase the temperature range. The paper also shows the possibility of creating high temperature chips based on them. To determine the possibility of using these chips in high temperature applications, there have been conducted some studies of the chips with the operation temperature 150°C. The results showed that are at this temperature the operation frequency falls by 10% on average, increasing of power consumption is insignificant. This indicates that the stocks in the project and manufacturing technology are enough.
Keywords: microprocessor, serial rapidio switch, spacewire switch, system on a chip, soi technology, intelligent well, high temperature electronics
Visitors: 10870

12. Terrain modelling in simulation training systems [№4 за 2015 год]
Authors: Mamrosenko K.A. (kirillam@ya.ru) - SRISA RAS, Ph.D; Reshetnikov V.N. (rvn_@mail.ru) - Center of Visualization and Satellite Information Technologies SRISA (Professor, Supervisor), Ph.D;
Abstract: The work presents the methods of the earth\'s surface and ground elements visualization and their usage in the aerospace industry. It also shows the features of the underlying surface modeling using satellite imagery, SRTM data. The authors consider how initial data is stored in the Tile Map Service and Keyhole Markup Language formats. Development of the earth remote sensing and data processing systems allowed using earth surface visualization techniques with satellite images as textures. For example, the GEOEYE-1 satellite makes it possible to obtain images with the resolution of 0.5 meters per pixel. In order to optimize performance the images are stored as tiles, which are small images of the same size, fragments of a big picture. The paper describes the method for increasing the visualization subsystem performance when displaying multiple objects of the same type. This method makes the sequent bypassing of all nodes of the model to clone and assembles a new model, which contains the required number of copies of the original model. An array of vertices of the original model is changed in order to place a sample model in the desired spatial position. An array of normals is copied without changes, as well as an array of texture coordinates. However, it is possible that the number of vertices in the model is bigger than the number of texture coordinates. In this case missing texture coordinates are replaced by the values from the next sample model, which can lead to incorrect texture mapping. To avoid such situations missing texture coordinates values are set to zero.
Keywords: tms, terrain, render, training simulation systems
Visitors: 10100

13. Ergonomic voice interface to control an anthropomorphic robot [№4 за 2015 год]
Authors: Mikhaylyuk M.V. (mix@niisi.ras.ru) - SRISA RAS, Ph.D; Torgashev M.A. (mtorg@mail.ru) - SRISA RAS, Ph.D; Omelchenko D.V. (denvikom@gmail.com) - SRISA RAS (Lead Programmer);
Abstract: Anthropomorphic robots (robots structurally similar to human beings) are widely used in various fields of scientific and economic activity. It is due to psychological factors, as well as to the fact, that many man-made objects are adapted to manipulate by human hands. However, ergonomic control of such robots is a serious and pressing problem. Standard computer interface devices (keyboard, mouse, joystick etc.) do not provide a sufficient level of ergonomics. Using control panels is also difficult because each control panel element usually affects one joint of the robot. Or alternatively, exoskeletons are often used for control. They are rigid articulated construction worn by an operator and match the organization of human body. An exoskeleton allows implementing the so called copying control mode when the robot will follow the operator’s exact movements. However, the exoskeleton is often difficult to use, either because of its great weight, or because it is impossible to use it in the suit. This paper proposes a voice interface to control a robot. The idea is that when the operator’s command is pronounced it is recognized and then in case of success the prepared script starts running, so the robot performs the specified action. Testing of the proposed methods and algorithms showed that it is possible to use them as an ergonomic interface of anthropomorphic robots\' voice control.
Keywords: virtual modeling, simulation and training complexes, anthropomorphic robot, voice interface
Visitors: 8995

14. The concept of mathematical and computer simulation of thermal processes in electronic systems [№4 за 2015 год]
Authors: Madera A.G. (alexmadera@mail.ru) - SRISA RAS, Ph.D;
Abstract: The article presents the concepts of electronic systems thermal design, as well as the requirements to mathematical and computer simulation of thermal processes in electronic systems. The author formulates the requirements to the mathematical models, which provide adequate simulation of thermal processes when thermal designing of electronic systems. According to these requirements, the mathematical models must be dynamic, nonlinear, stochastic in intervals, consider the effect of thermal feedback (interaction between thermal and electrical modes in the electronic system) as well as statistical technology spread of electric and thermal defining parameters of elements in the electronic system. The obtained equations of the mathematical model for thermal processes in electronic systems meet the stated requirements. They are a system of the time-dependent, nonlinear, stochastic interval ordinary first-order derivatives. The developed concepts and mathematical models may be taken as the base of the multi-functional software package for thermal design of electronic systems.
Keywords: thermal feedback, stochastic, interval, nonlinear, dynamic, computer modeling, mathematical model, thermal process, electronic system
Visitors: 11542

15. High performance RapidIO block to create multi-core microprocessors with RapidIO virtual links [№4 за 2015 год]
Authors: Kozlov N.A. (kozlov_n@cs.niisi.ras.ru) - SRISA RAS (Junior Researcher), ; Bobkov S.G. (bobkov@cs.niisi.ras.ru) - Federal State Institution "Scientific Research Institute for System Analysis of the Russian Academy of Sciences" (SRISA RAS) (Professor, Deputy Director);
Abstract: One of the main methods of improving computing systems performance is the development of multi-core and multiprocessor systems. Modern multiprocessor systems are usually based on the communication environment. The most popular high-performance network environments are: HyperTransport, PCI Express, ASI, RapidIO, VXS, StarFabric, 10GB Ethernet, InfiniBand, Myrinet. A communication environment is created on chip level, module level, inter-module level and inter-machine level. The RapidIO standard is focused on creating the last three environments and thus we can unify a multiprocessor system. For further unification we need to develop the module of the exchange between the microprocessor’s core and the external environment in order to improve the reliability and reduce labor effort. The article discusses a unified adapter block (RIO-AXI) for a transition from on-chip AXI bus to the external RapidIO bus to create as multicore processors as well as switches with RapidIO environment. Chip production technology improvement and increase in the number of transistors in the chips until 2003 has led to permanent growth of microprocessor frequency up to 3–4 GHz. However, further increase in frequency has caused a substantial increase of power consumption and heat dissipation, which is impossible to withdraw from crystals using standard means. Increased productivity has been achieved by expenses on complicated architecture of microprocessor cores, additional functions, coprocessors and, most importantly, by creating multi-core microprocessors and multiprocessor systems. Therefore, a high-performance communication environment is emphasized. RapidIO communication environment is one of the most promising. The RapidIO standard is designed specifically to meet critical requirements of real-time applications, which are: low latency, determinism, reliability and scalability, reduction of power consumption, size and weight.
Keywords: virtual channel, multi-core processor, RapidIO, switch
Visitors: 8493

16. MPI support in Linux OS for RapidIO-based multiprocessor computer systems [№4 за 2015 год]
Author: Kuleshov A.S. (rndfax@cs.niisi.ras.ru) - SRISA RAS;
Abstract: Solving calculus problems using multiprocessor computing systems refers to a certain data transfer programming interface. The most well-known interface is the message passing interface (MPI), which defines the standard of API functions to transfer data between nodes for end-user programs. This paper presents MPI implementation for SoC 1890VM6IA developed in SRISA RAS. It is a base for multipurpose multiprocessor computing systems, nodes of which can be connected through high speed RapidIO channels. To implement it we took MPICH, which is the most widespread library of MPI, and Linux kernel that proved itself as the best solution for supercomputing. MPICH basic configuration includes different network modules, which implement API for specific transport drivers: tcp (Ethernet TCP/IP), ib (Infiniband), mx (Myrinet eXpress) and others. Using network module mx, which is well-fitting for RapidIO architecture, gave the opportunity to focus only on the development of the Linux driver for RapidIO controller. The features of RapidIO controller of 1890VM6IA, MPI interface and the idea of using RapidIO channel for purposes other than MPI in Linux, allowed implementing a message passing stack with the some degree of universality using RapidIO channel without unnecessary data copying. In the end of the paper there are the results of different MPI test programs such as NAS Parallel Benchmarks and OSU Micro-Benchmarks for 4 nodes using the RapidIO channels. The author made a conclusion and summarized the results using RapidIO controller of 1890VM6IA SoC as a transport facility for MPI.
Keywords: linux, mpich, MPI, RapidIO
Visitors: 8847

17. Textured planet model visualization system for space experiments simulation [№4 за 2015 год]
Author: Timokhin P.Yu. (webpismo@yahoo.de) - SRISA RAS (Research Associate);
Abstract: A modern tool to train space experiments related to the planet observation from a circumplanetary orbit is real-time visualization of virtual 3D planet model with detailed textures based on satellite images. Studies show that the size and memory space required for such textures significantly exceeds the capability of modern graphics accelerators (extra-large textures). This does not allow loading such textures in video memory completely and their hardware processing. This article describes a technology and software implementation which allows real-time visualizing of extra-large planet textures divided into small pieces of an equal size (pages). The key idea of the proposed technology is to reduce time costs for visible pages selection using GPU parallelizing computations by means of developed shader programs. The technology provides texture detailing of the simulated planet surface to fit the screen resolution. It also allows real-time visualizing multiple extra-large textures with different initial resolution on a virtual planet surface. The proposed technology is implemented in the planet texture visualization software package, which can be integrated in an external application. The created software package is also equipped with a module which calculates a observer\'s orbital position, as well as a user interface to control an observer\'s flight. The paper considers testing of the implemented software package as a part of 3D virtual scene visualization system. The results show that it meets all the requirements. The developed software package can be used both to improve features of existing simulators for space scientific experiments and to build new promising ones.
Keywords: real time, extra-large texture, virtual planet, space experiment, render
Visitors: 9051

18. Collision detection for bounding spheres and rectangular parallelepipeds in 3d modeling systems [№4 за 2015 год]
Author: Trushin A.M. (leha_trushin@mail.ru) - SRISA RAS (Research Associate);
Abstract: Virtual objects in 3D modeling systems may collide with each other. Collision detection is an integral part of any physical engine. The speed of calculation is crucial for physical engines. In real -time mode one simulation frame calculations should not exceed 40 ms to visualize at least 25 frames per second. Therefore, there is a need in development of fast and efficient algorithms for the dynamics calculation system and for collision detection in particular. Collision detection of complex shape\'s objects is a difficult task, which has a high computational complexity. Therefore, a method using bounding volumes is widely used. In this case, virtual objects forms are described by as different geometric primitives, and the problem of objects\' collision detection is reduced to the collision detection of their bounding volumes. Such primitives as rectangular parallelepipeds (boxes) and spheres became widespread. Collision detection algorithms may be a priori and a posteriori. A priori algorithms predict collisions of bodies, and a posteriori algorithms detect collisions after actual intersections of the objects. In general, a priori algorithms have much higher computational complexity due to the greater amount of input data. In this regard, physics engines oriented on real-time dynamics modeling basically use a posteriori collision detection algorithms. This work is devoted to the development of fast and efficient algorithms for a posteriori sphere-sphere and sphere-box collision detection.
Keywords: dynamics modeling, approximating containers, collision detection
Visitors: 6582

19. Automated design flow of self-timed pipelines using EDA tools [№4 за 2015 год]
Author: Vlasov A.O. (vlaalex@cs.niisi.ras.ru) - SRISA RAS (Head of Sector); Surkov A.V. (surkov@cs.niisi.ras.ru) - SRISA RAS;
Abstract: The self-timed logic doesn’t use the clock signal and operates in a whole range of environmental parameters, where transistors still can switch. But there are some weak points due to which the self-timed logic is not used along with the traditional synchronous logic. The design of self-timed schemes has two major approaches: signal transition graph (STG) design and the so-called burst-mode technique. The first approach provides the fastest and the smallest schemes, but due to high complexity of working with graphs the scheme size is limited by a number of elements. The burst-mode technique is the best to design big logical blocks, but requires the four-phase dual rail protocol, which leads to logic doubling in the resulting scheme. As a result, the burst-mode logical blocks operate at a half speed and occupy much higher area. Nevertheless, the self-timed logic may still be used in applications, which require extended temperature range and/or supply voltage variation, where the synchronous logic is mostly unavailable. The major disadvantage of all self-timed techniques is the absence of the automated design flow, so self-timed schemes are mostly implemented manually. To reduce the design time we have developed the design flow based on previously created synchronous prototypes. It includes the synthesis, layout and functional verification via modeling.
Keywords: encounter, design compiler, CAD system, pipelines, self-timed, asynchronous
Visitors: 8351

20. Kalman filter as a method of secondary processing of GLONASS system information [№4 за 2015 год]
Authors: Bolodurina I.P. (prmat@mail.osu.ru) - Orenburg State University, (Professor, Head of Department), Ph.D; Nugumanova A.A. (alaymir@mail.ru) - Orenburg State University; Reshetnikov V.N. (rvn_@mail.ru) - Center of Visualization and Satellite Information Technologies SRISA (Professor, Supervisor), Ph.D;
Abstract: The paper discusses methods of positioning a land vehicle based on secondary processing of information from GLONASS satellite navigation system. Vehicle positioning accuracy depends on many factors including the method of finding the coordinates. It is one of the indicators of vehicle effective control. Therefore, one can quickly control the vehicles movement, their actual value and the fuel consumption accuracy. The publications devoted to the problem of dynamic objects positioning, as a rule, do not describe mathematical models and algorithms, they just display the final results. In fact there are no sources, which can clearly describe an algorithmic scheme suitable for developing software. This is due to the fact that the software is a trade secret or an intellectual property. In contrast to existing methods of mobile objects control and monitoring, the authors have developed a Kalman filter algorithm step by step. It increases the positioning accuracy of a land vehicle. The algorithm is a basis for application software, which determines the land object location in real-time and post-processing mode. The paper discusses the cases of mobile and stationary state of a land vehicle. Each case shows using Kalman filter algorithm and the least squares method to enhance positioning accuracy. There is a comparative characteristics of these methods on the basis of selected statistical criteria. The authors set the most accurate method of positioning mobile and stationary objects.
Keywords: method of least squares, Kalman filter, vehicle, glonass, positioning
Visitors: 14142

← Preview | 1 | 2 | 3 | 4 | Next →