full factorial experiment in study of control block operation of intersection passage by unmanned vehicle

A statistical evaluation of the control unit computer model for multi-agent robotic systems is carried out. An algorithm is considered that realizes the achievement of a single goal by robots-agents – to cross the intersection at the maximum speed. Let such algorithm be called an "algorithm for conflict solutions". The paper presents the study of the dependence of the vehicle appearance probability effect from each intersection side on the maximum speed of the intersection crossing. The efficiency of the applied algorithm is evaluated at the intersections of various types, for example: four-way intersection, T-junction, Y-shaped intersection. The efficiency estimation was performed by a full factorial experiment on the study of the "algorithm for conflict solutions" operation by the unmanned vehicle control unit. The basic idea of a full factorial experiment is that the researched object is considered as a "black box". Thus, the effect of input parameters (factors) on the response function is studied. The main benefit of full factorial experiment application is the construction of an experiment planning matrix that contains all possible combinations of factors. This approach gave the possibility to identify those combinations of factors that have the greatest and the least effect on the response function. The received results testify to the possibility of the algorithm for conflict solutions operation for intersection of various types of roads. The algorithm shortcomings were revealed. In future it is expected to use the obtained results to assess the information security of the multi-agent robotic system with decentralized and centralized control.