Search

Search Results

• MACHINE LEARNING MODEL OF SWARM EVASION FROM THE INFLUENCE OF ANTAGONISTIC ENVIRONMENT

  V. К. Abrosimov, G.А. Dolgov, Е. S. Mikhailova

  6-19

  2025-04-27

  Abstract ▼

  One of the priority areas of group control theory for the near future is swarm control of groups of
  small unmanned aerial vehicles - micro-, mini- and nano-classes, performing a collective task under enemy
  influence. Here, two antagonistic strategies collide - minimization of losses from the point of view of
  the attacking swarm and maximization of such losses from the point of view of the defense system. Research objective: development of an approach to solving a practical problem - penetration of a swarm of
  unmanned aerial vehicles into an object protected by a defense system. The objectives of the study were to
  analyze the characteristics of the factors influencing the processes of detection, tracking, recognition of
  swarm intentions by the defense system and the development of a machine learning model for creating
  spatio-temporal formations that minimize the number of swarm elements affected by the defense system.
  The main parameters of the defense system are the detection range and duration of swarm recognition, the
  time to make a decision on the actions of the swarm, the size of the zone of destruction of defense means.
  The method of machine learning on convolutional neural networks with reinforcement was chosen as the
  research method. The counteraction effect against the defense system is created due to the swarm's dynamics;
  it can actively maneuver, creating spatio-temporal maneuvers during the mission. To simulate the
  "Swarm vs. Defense System" situation, a swarm agent (a neural network with a transformer architecture
  that initiates swarm formations) and a defense system agent are introduced that recognizes the swarm and
  attacks it, creating a zone of destruction in the conventional center of mass of the swarm. The swarm is
  guided by a stochastic rule, asking the defense system (environment) to react to its maneuver. The environment
  responds by attacking the swarm, creating a damaging factor at the point where the swarm or the
  main part of the swarm is expected to be. The reward of the swarm strategy is the number of undestroyed
  objects under the conditions of constraints; for the defense system, this "reward" acts as a "punishment".
  An interesting phenomenon was established in the process of machine learning: each swarm element,
  remaining within a given space and implementing the biological principles of swarm control without a
  Leader, independently evades the area of destruction, which together creates a random spatio-temporal
  formation for defense means with minimal losses of swarm elements. Thus, using the method of machine
  learning with reinforcement, a model was created that allows varying the behavior of the swarm and synthesizing
  spatio-temporal formations that complicate detection, tracking, recognition of intentions and
  decision-making on the impact of the defense system on a swarm of attacking small unmanned aerial vehicles,
  as well as significantly reducing their losses.

• ON THE USE OF SIMILARITY THEORY TO ASSESS THE DYNAMICS OF CLUSTERS OF SUBJECTS OF INTEREST ON THE GROUND

  V.К. Abrosimov, S.М. Lapin

  2024-04-15

  Abstract ▼

  The experience of modern combat operations has initiated the high relevance of the tasks of airdelivered
  assessing the dynamics of changes in time characteristics of groups (clusters) of objects of
  interest on the ground. The active development of unmanned aviation, including within groups, provides
  new opportunities for periodic monitoring of the area with the solution of problems of detection and
  recognition of clusters of objects of interest in dynamics. The article analyses the possibility of using the
  theory of resemblance to solve the problems of assessing the similarity of types of weapons, military and
  special equipment by the nature of distribution in various clusters, including in various geographical
  conditions. It is shown that the dynamics of objects can be established by regular monitoring of the terrain
  with the estimation of various measures of similarity and difference for clusters. At the same time,
  the applicability of well-established statistical methods of biodiversity research developed in biology to
  assess the diversity of population, their complexity, similarity, relationships, etc. is proved. The characteristics
  of the species diversity of the most important deterministic clusters of troops and equipment of
  NATO countries are given. The efficiency of the proposed approach demonstrated by the example of
  aerial reconnaissance of a conditional area with recognition of the dynamics of five types of clusters,
  including various types of military personnel, personnel and engineering equipment. General recommendations
  for conducting appropriate assessments and decision-making are given.The following basic similarity measurements are recommended for use: Jacquard similarity coefficients to determine the
  similarity level of clusters by their constituent types of VVST samples (cars, tanks, guns, armored vehicles,
  etc.), the Margalef index to determine the number of types of VVST in the total number of VVST
  units in the cluster, the generalized Shannon diversity mass to assess the diversity of species in a cluster,
  the Sorensen-Chekanovsky coefficient is used to determine the degree of occurrence of the selected type
  of samples in the cluster. It is advisable to use the obtained results in multi-criteria tasks of preflight and
  operational planning of group operations of unmanned aerial vehicles in the interests of monitoring the
  controlled territory, taking into account the required schedule for obtaining reliable information.

• EFFICIENCY OF THE COLLECTIVE MISSION BY A MULTIGENT HETEROGENEOUS GROUP

  V.K. Abrosimov, А.N. Sedov

  2023-04-10

  Abstract ▼

  The article deals with study оf the criteria, parameters and performance indicators for the implementation
  of the collective mission of extended space monitoring by a group that includes control
  objects with different functionality. The most urgent military-technical task of recent times – monitoring
  of extended spaces- is considered as an example of a mission. A feature of the article is the assumption
  about of hazardous zones in the monitoring area. Any counteraction from potentially critical
  objects may significantly limit the capabilities and even disable the monitoring tools. This proviso
  affects for implementation of the mission by the group and leads to a revision of the group collective
  strategies and individual strategies for monitoring tools, their movement routes, and decision-making
  algorithms and etc. Monitoring tools are modeled as intelligent agents trained by the "collectivism"
  paradigm. It provides a common resource of situational awareness of the group, the organization of negotiations between agents and mutual assistance to each other in the event of problem situations.
  A general approach has been developed to evaluate the effectiveness of solving collective problems
  by a heterogeneous group of agents, taking into account the objective function, resource intensity of
  the entire mission and organizing active interaction between agents. It is proposed that the mission
  efficiency characteristic expressed as a weighted sum of the normalized indicators of the parameters
  of the agents' functions with the weights "significance of the function for the mission" and "value of
  the object". It is shown that the loss of efficiency in the performance of a collective mission in the
  event of problematic situations with an agent (active enemy action, breakdown, lack of the necessary
  functionality, resources, etc.) can be compensated by the required functionality of other agents of the
  group with the appropriate reconfiguration of tasks.