Application of minimax evaluation procedures for expansion of metrological standard on noised measurements processing

The paper deals with an application of minimax evaluation procedures for expansion of metrological standards in the case of information uncertainty about errors in measurements. It is concluded that the discussed approach for evaluation of noised experimental data does not contradict the procedures of existing standards, but can complement them when the measurement error´s statistical characteristics are unknown. The approach gives additional and practical information: a full representation of the set of admissible parameters for the describing function; a full description of admissible values of the process to be evaluated (the tube). Elaborated algorithms have rather simple numerical realisation