Neural networks models and their training algorithms on a central computer with reference to a previously developed distributed sensor network are considered. The requirements for its intelligent node are formulated. Also the node's structure is offered which realises such intelligent functions, as sensor and other measuring channel components drift prediction using remote reprogramming.
In this paper the data acquisition module based on sigma-delta analog to digital converter (ADC) and MCS-51 compatible microcontroller with dynamic remote reprogramming is presented. The main feature of presented module is in integral nonlinearity of ADC correction. The error of proposed correction is considered in this paper.
There is proposed method of control which could be implemented for heaters situated across electrodes of the main thermocouple and provides stabilization profile of temperature field. The proposed method has small requirements to memory space and computing power so it could be implemented using 8-bit microcontrollers.
This paper proposes a method for fuzzy data processing based on Mamdani's fuzzy inference method. The implementation of this method is divided into phases of training and exploitation, which reduces the number of operations during fuzzy data processing and improves its performance. The analysis of the time complexity of the proposed method shows the reduction in run-time processing of fuzzy data four times comparatively to the time of the realization of fuzzy inference by classical Mamdani's mechanism.