ABSTRACTS
A Method for Ultralight UAV Orientation with Rare
Updates of Its Location Data
K. S. Amelin
Saint Petersburg State University
konstantinamelin@gmail.ru
Key words: unmanned aerial vehicle,
navigation system, randomized algorithm, object recognition, flight
optimization.
In recent years,
ultralight UAVs are increasingly used to meet the challenge of area monitoring,
both in military and civilian sectors. The main criterion that guarantees the
task implementation is the accuracy of orientation. Usually, the basic
equipment for locating is a GLONASS/GPS receiver; the signal from which may be
received with noise, delays or could be absent. In this paper, a new ultralight
UAVs orientation method is proposed; it is based on detecting a priori specified
landmarks while processing the data obtained from the onboard photo camera and
successively sweeping the landmarks. The algorithm for generating the landmarks
data of the area, action scenarios for UAVs, and a randomized algorithm to
optimize the motion of the UAV are described. An example of the ultralight UAV
architecture is considered to show the applicability of the proposed method.
Bibliogr.: 11 refs.
Scenario Approach and Feedbacks Using PI-controller Under
Idealized Market Conditions
A. D. Kornivetc
Saint Petersburg State University
alexandra_91@inbox.ru
Key words: financial mathematics, scenario
approach, adaptive strategy, feedbacks, PI-controller.
A number of recent
papers, the apparatus of financial mathematics was used to elaborate new
approaches to the analysis of stock exchange. In particular, B. Barmish and
co-authors exploit the PI-controller approach. The main attention is paid to
the idealized market conditions and prices generated by a non-trivial Geometric
Brownian Motion. Under these conditions, a combination of static and dynamic
linear feedbacks leads to the positivity of the robust mathematical
expectation. Such a strategy does not guarantee a profit; however it shows
itself efficient in simulations. Additional assumptions on the model
uncertainties are required; specifically, those related to the range of
possible variations of the parameters of the moving average model. In this
paper, we provide a survey of the stock exchange trading strategy mentioned above,
together with the scenario approach. This latter was chosen purposely, since
approach can be used to optimize the choice of the feedback coefficients.
Bibliogr.: 8 refs.
Generation of Graphs with Prespecified Sequences of
Degrees of Order Two and the Isomorphism Detection Problem
B. F. Mel’nikov
Samara State University
bormel@rambler.ru
E. F. Saifullina
Togliatti State University
elena-fairy@yandex.ru
Key words: graphs, sequence of degrees,
random generation, multiheuristic approach, isomorphism.
In this paper, a
multi-heuristic method, a version of incomplete branch-and-bound method is
proposed and applied to the random generation of graphs with a given vector of
degrees. We also consider various graph invariant comparison sequences
understood as heuristic algorithms for checking the isomorphism. Finally, we
provide the results of numerical experiments targeted at the quantitative
comparison of efficiency of using various invariant comparison sequences.
Bibliogr.: 14 refs.
Aggregation of Measurement Systems Based on Kalman
Filtering
V. M. Ponyatskiy
KBP, Tula
kbkedr@tula.net, pwmru@rambler.ru
Key words: Kalman filter system, automatic
tracking, aggregation.
The possibilities of
using the Kalman filter to the aggregation of measuring systems are considered.
A mathematical model of the two measuring systems with Kalman filter is
developed and tested.
Bibliogr.: 5 refs.
A Method for Evaluating the Performance of Algorithms
of DNA Inter-String Distance Detection
R. I. Semenov
Togliatti State University
romansemenov3@gmail.com
Key words: random variables, DNA string,
string difference, algorithm performance assessment.
Discoveries of modern
biology and biochemistry increasingly requires processing large amounts of
data. One of the promising areas of application of information technology is
analyzing the transcribed DNA strings. By comparing the DNA strings, it is
possible to establish (with certain probability) the relationship between the
individuals and the proximity between the analyzed species. In this paper we
describe a method for estimating the performance of algorithms that compute the
distance between the DNA strings; application of this method to the evaluation of
the existing algorithms is also considered.
Bibliogr.: 12 refs.
