Abstract
The problems of developing new generation knowledge acquisition systems are faced. A crucial role of measurement in acquiring knowledge for intelligent systems is shown. A comparative analysis of classical measurements and expert estimates is made. Some nonclassical measurement concepts based on granular ontologies and measurement information granulation are presented. The classification of non-traditional measurements is given. The Russian scientific tradition of considering measurement as a cognitive process is discussed. The concept of Cognitive Measurement based on Cognitive Sensors is introduced. Here Cognitive Measurement is viewed as a two-leveled granulation process where the lower level is responsible for obtaining fine-grained data by artificial sensor system, and the higher level transforms it into coarse-grained information related to a pragmatic scale of normative linguistic values such as "norm", "nearly norm", "out of norm", "far from norm" etc. Two types of Cognitive Sensors with appropriate pragmatics are introduced. A bilattice-based interpretation of multi-sensor data fusion is proposed.