| Multisensor information fusion is a new technique developed in recent year. Through the integration of redundant, complementary or even real-time information provided by multiple sensors, systems can obtain a more accurate and more precise information.Multi-sensor systems usually have a cross-sensitivity problem, the sensor static characteristic affected not only by a certain environment parameters, and sometimes even by a variety of non-target parameters, resulting in lower measurement accuracy.In view mutual disturbance question between pressure and temperature in pressure and temperature measurement systems, a method of multi-sensor intersecting sensitivity restrait based on BP neuralnetwork is presented in this paper. By comparing the data t do not use neural network with the data use it, the results show that the method can greatly reduce the cross-interference with each other, improve measuring precision and come to the anticipative effect. In the paper the latest STM32F101C8 micro-controller based on CORTEX-M3 of ST is utilized to achieve the data fusion between pressure and temperature sensor based on BP network.This paper detailedly expatiated on the design proposal for the intellectualized sensor system, which mainly includes the following aspect:sensor circuit design principle, sensor measurement system hardware design, sensor intellectualization software design. In the software design uses real-time operating systemμC/OS-Ⅱto achieve real-time scheduling of tasks, complete the data acquisition, the data conversion, the data processing, exception alarm, display and communication functions.This topic combine theory research with practice exploiture, the successful realization of the BP network in STM32F101C8 by combining with the latest embedded system development platform. In this intellectualized sensor achieve data fusion between pressure and temperature sensor based on BP network.the cross-sensitivity to be well constrained,to achieve good results. |
PDF Full Text Request