Uncertainty-management-network-based dynamic sensor model

Abstract

The raw data obtained by physical sensors are initially modeled using fuzzy numbers which are then processed by the subsequent uncertainty management network (UMN) which is a new paradigm in propagating uncertainties through a sensor system model. The UMN partitions the processing blocks of the sensor system into a tree-like network structure of basic processing nodes which perform elementary arithmetic, logical, aggregation, or branching operations interconnected using multiple information propagation channels. The UMN allows the dynamic modelling of sensor systems by providing a confidence measure for the output of the sensor system which incorporates the changing conditions of the environment as well as the changes occurring within the sensor system itself. An example of an UMN-based vision system is illustrated to clarify the idea and the concepts.<>