On soft measurements and data mining based on granular pragmatics, multi-valued and fuzzy logics

The concept of Soft Measurements as a natural Counterpart of Soft Computing is discussed. The differences between hard and soft measurements are shown. To do this main features of Wireless Sensor Networks are considered. The tight connections between soft measurements and mining data from sensor networks are revealed. A new approach to soft measurements based on distributed cognition and information granularity is proposed. Such fundamental granulation concepts as granular algebraic system and fuzzy granular logical matrix are introduced, the notion of logical pragmatics is clarified. These basic concepts underlie the construction of logical granules. Two important sensor types for intelligent sensor networks - Vassiliev´s sensor and Belnap´s sensor - are suggested, the representation of the latter by logical lattice and minimal bilattice is considered. A bilattice-based model of sensors communications and negotiations in sensor network is developed. Interpretations of basic granular pragmatic truth values for two communicating sensors are suggested. The map for dealing with different sensor data inconsistencies is constructed. The transition from crisp to fuzzy Belnap´s sensors equipped by fuzzy pragmatics is justified. The fuzzy dialogue model in sensor network is developed.