Research On Data-Collection Preprocessing In Emergency Communication Monitoring And Controlling System

Emergency communication refers to a means of communication and methods, which are used to protect the emergency rescue, relief and the necessary communications when human or natural emergency situations occured and communications equipments are damaged. Emergency communication monitoring and controlling system is designed to monitor status of system in real time, to detect system fault and protect the normal operation of emergency communication system. However, due to numerous equipment and much large-scale network in emergency communication system, limited bandwidth communication networks are bringing pressure on mass data collected by the data acquisition system, while complex and diverse data indicators affect the efficiency of fault diagnosis. Therefore, based on monitoring and controlling system, it is important to deal with the problems, namely how to conduct the relevant treatment on the collected data in order to achieve reduction in the amount of information, reduce network load, decrease the occupancy of the communication resources under the premise of without lowering the rate of fault diagnosis.It is researched data preprocessing-related technologies in the background of emergency communication system, combined with hierarchical and distributed multi-domain monitoring architecture. The results include the following aspects:(1)Based on structural characteristics of hierarchical and distributed multi-domain monitoring and controlling system and studying a variety of data preprocessing methods, hierarchical and distributed multi-domain data preprocessing system model is proposed. In this model, all nodes are divided into multiple domains, thus the load of the master node is small,the system can respond rapidly and has good scalability. On this basis, the data reduction and data integration has been achieved.(2)In order to transmit data which is more targeted to the fault diagnosis end, it is presented attribute selection mechanism which based on signal feedback of fault diagnosis end, it is established rule base which is desired by attribute selection. This mechanism effectively reduces the number of indicators, reduces the bandwidth of possession, will enhance the effectiveness of fault diagnosis.(3)To solve the overlap of information between data indicators, data redundancy and other issues, it is explored in depth the principles of principal component analysis, characteristics and processes, at the same time, the principal component analysis method is applied to monitor the emergency communication system. Simple data is obtained when real-time monitoring data is to be projected to the feature subspace which formed from model training. In order to comprehensively and objectively train data, multi-domain and multi-database integration has been achieved.Based on the results mentioned above, system simulation platform has been built which verifed the efficiency of selection methods and principal component analysis method. The experimental results show that the methods of attribute selection and principal component analysis are efficient to eliminate redundant information between the indicators and to reduce the amount of information.