UNMANNED VEHICLE CONTROL SYSTEM BASED ON FUZZY CLUSTERING. PART 2. FUZZY CLUSTERING AND SOFTWARE IMPLEMENTATION
This article continues the description of the control system for ground unmanned vehicles as part of the integration of a phenomenological approach to modeling the behavior of agents and methods of fuzzy clustering in order to improve the quality of decisions. As a result, adaptive fuzzy clustering methods provide support for adaptive ground unmanned vehicles control, which minimizes the risks of accidents (emergencies involving ground unmanned vehicles) and maximizes traffic (total output stream) in conditions of heavy traffic. The second part is devoted to the description of the developed fuzzy clustering algorithm, software implementation and experiments. As a result, within the framework of the developed model of ground unmanned vehicles movement, fuzzy clustering methods are used to ensure the procedure for choosing the most preferable (least dense) lane in conditions of heavy traffic and to support continuous information exchange between ground unmanned vehicles. The software implementation of the developed simulation model in the AnyLogic environment was performed and numerical experiments aimed at analyzing scenarios of the development of the road situation with the participation of the ground unmanned vehicles ensemble were carried out. Various behavioral scenarios of the developed ground unmanned vehicles control system were investigated, and agent clustering was performed for each scenario under consideration. As a result of numerical experiments, the effectiveness of using the proposed fuzzy clustering procedure to assess the density of the road flow and adaptive control and maneuvering of the ground unmanned vehicles is confirmed.