Research On Location Algorithms For Interval Sectional And LOS Nodes Cooperation In Coal Mine

The environment of coal mines is complex and variable,there are many disturbance factors affecting the transmission of electromagnetic waves.The accuracy of positioning of underground personnel is greatly challenged.Accidents in underground accidents are difficult to avoid,accurate and reliable location service is the technical support for search and rescue in the face of disasters.Mine IoT technology promotes the use of WiFi in the underground coal mine positioning.Because of the simple deployment,low cost and wide coverage,WiFi is one of the hot research topics to locating in complex space.At present,underground electromagnetic wave signals are seriously interfered by multipath effects and Non Line of Sight paths,resulting in low positioning accuracy.In order to overcome the transmission interference of WiFi in complex environment and improve the positioning accuracy,it is of great theoretical significance and practical value to further study the positioning technology based on WiFi personnel.Based on the analysis of the actual positioning environment and the study of machine learning algorithms,In-depth analysis of the advantages of signal strength value RSSI and channel state information CSI in WiFi communication technology.This thesis proposes the Sectional Interval and LOS Node Cooperation Localization Algorithm in Coal Mine.The Channel State Information(CSI)in the WiFi communication technology is fused with orthogonal frequency division multiplexing(OFDM)technology,which can overcome the interference of multipath propagation.The algorithm first extracts the kurtosis and delay characteristics of CSI as SVM input features,and uses SVM feature classification technology to accurately identify the lineof-sight(LOS)path information under multipath effect.Then,using the identified signal strength value RSSI under the LOS path to construct an accurate log-normal shadow attenuation Shadowing model.Aiming at the problem of serious signal attenuation during long-distance regional positioning of coal mine,the learning vector quantization LVQ algorithm is used to segment the long-distance positioning region.Then,according to the segmentation parameter,the interval to which the node to be tested belongs is determined,and the located nodes is regarded as a “virtual reference node”.Determining whether the virtual reference node is available according to signal strength attenuation and data stability.If the condition is better than the actual reference node,the virtual reference node in the interval is used to replace the actual reference node of the long-distance interval,so as to realize cooperative positioning of different types of nodes,and use the selected cooperative node information to realize initial positioning under the three-sided weighted positioning algorithm.This part can reduce the influence of large energy attenuation of the reference node signal on the positioning at a long distance,and enhance the robustness of the algorithm.Finally,in order to further reduce the positioning error,the initial positioning result is corrected by using the coordinate correction algorithm to calculate the node position more accurately.Through the actual roadway scene test,the maximum error of the algorithm is 3m,and the average positioning error is 1.5m,which basically meets the requirements of underground personnel positioning.