The Research On Positioning Algorithm Of Adaptive Fuzzy Neural Inference Based On RSSI

The geological structure of coal in China is complex,and most of the coal mining work is carried out under the mine.Due to the poor production environment provided by the coal mine,the continuous development of the underground positioning technology has become the prerequisite to ensure the safety of coal mine production.In recent years,the underground positioning technology has gradually developed from wired communication to wireless communication.Among them,wireless sensor networks is a kind of wireless communication technology.It has the characteristics of low power consumption,low cost,convenient deployment,wide coverage,strong robustness,etc.WSN can cooperate with various portable mobile hardware to achieve positioning,which is one of the important means to achieve down-hole positioning.However,the underground environment structure is complex,the underground workers and the running mechanical equipment always interfere with the transmission of wireless signals,which brings many difficulties to the effective realization of underground positioning.Under the support of NSFC and Inner Mongolia Science and technology program,this paper studies the localization algorithm of wireless sensor network in coal mine.In order to reduce the influence of RSSI signal,save the labor of fingerprint database and improve the positioning accuracy,this paper successively uses fuzzy reasoning,Bayesian probability estimation,adaptive fuzzy neural network,clustering analysis and HMM and other theories.The results show that the algorithm proposed in this paper improves the positioning effect of fingerprint matching in coal mine to a certain extent.The main contents of this paper can be summarized as follows:(1)A fingerprint matching algorithm based on Bayesian fuzzy probability is proposed in the coal mine.Considering the time-varying characteristics of communication environment in coal mine,the RSSI reliability mechanism based on calibration node is established.The posteriori probability of RSSI signal at each reference point in fingerprint map is calculated by Bayesian probability inference formula.The reliability mechanism and Bayesian posterior probability are input into the fuzzy inference system,and the weight is calculated according to the designed fuzzy rules.The RSSI reliability of the calibration node added to the algorithm is equivalent to incorporating the environmental factors,whichreduces the impact of environmental changes on the location of fingerprint matching based on Bayesian fuzzy probability,and makes the algorithm robust to a certain extent.(2)A fingerprint matching algorithm model based on ANFIS is established in the coal mine.The ANFIS positioning model uses its own learning mechanism to automatically learn and adjust the parameters of the membership function of the fuzzy system,and finally establishes a fingerprint matching positioning model more in line with the coal mine environment.Using this model to realize the positioning in the coal mine not only has the advantages of simple operation and low calculation complexity,but also improves the accuracy of fingerprint matching positioning.(3)On the basis of the reasonable division of coal mine underground location area,a dynamic updating algorithm of underground fingerprint based on HMM is proposed.It is time-consuming and labor-consuming to set up fingerprint database for the whole tunnel area because of the large coverage of underground tunnel.Aiming at the positioning sub area with significant environmental change,this paper uses the sensor nodes worn by the underground workers to collect the RSSI sequence in real time.The RSSI sequence collected is used as the observation sequence of HMM,and the corresponding position information is decoded and calculated.The real-time RSSI sequence and decoded position coordinates are used as new fingerprint information to update the database.The dynamic updating algorithm based on HMM can save labor and improve the reliability of fingerprint matching and positioning.(4)A positioning system belonging to the coal mine is built.Several different positioning algorithms are realized by using the positioning system,and the positioning performance of each algorithm is analyzed and summarized.