Research On Ad-hoc Communication Signal Monitoring And Node Association Analysis

As a communication network that can be built quickly without fixed infrastructure,Ad-hoc network,with strong flexibility,can meet the needs of emergency communications and other special occasions.With the development of communication technology in recent years,Ad-hoc network gradually develops the characteristics of high confidentiality,large number of nodes,random frequency and time hopping,etc.The centerless,flat and flexible networking and clustering of the network bring some challenges to the electromagnetic spectrum management and spectrum security analysis of Ad-hoc,so the monitoring of Ad-hoc communication signals and the analysis of node association are of great significance.The content of this paper covered the whole process of communication signal information mining,including signal monitoring and receiving,signal measurement and sorting and node association analysis.The follow-up analysis is based on Real-time high-speed pulse monitoring according to the physical layer characteristics of Ad-hoc signal.Parameter measurement and sorting are the premise of subsequent node association analysis.This paper proposed two sorting methods in special circumstances as a supplement to the traditional sorting.Non-cooperative node association analysis only used pulsed TOA and TOE information,combined with Ad-hoc communication protocol or other prior information to complete the association between signals and the communication association analysis between nodes.This paper proposed two association methods in two different circumstances.This paper mainly focused on the following work of non-cooperative information mining of Ad-hoc network:(1)In terms of signal monitoring and receiving,the digital receiving algorithm of Ad-hoc communication signal with specific frequency hopping and time hopping characteristics was designed and implemented.According to the signal characteristics,the8-channel 50% overlapping digital channelization receiving algorithm and cumulant signal detection algorithm were designed,and the algorithm was implemented on FPGA and transplanted to USRP 2955 platform to complete the system joint tuning.(2)In terms of signal measurement and sorting,two sorting methods based on time hopping information and Doppler frequency shift information were proposed as a supplement to the traditional sorting.One is the intra-frame pulse sorting algorithm based on the specific time-hopping model.Relying on the special time-hopping relationship of intra-frame pulses,an intra-frame multi-pulse search algorithm was designed,which can achieve approximately 100% search accuracy in the absence of measurement noise.At the same time,it still has high performance under different signal density and measurement noise.The other is the sorting method based on Doppler shift.Considering that the signals emitted by aircraft with different speeds will have Doppler frequency offset from0 to 5k Hz and the traditional non-cooperative carrier frequency measurement algorithm cannot reach the frequency measurement accuracy,a frequency measurement algorithm based on neural network algorithm was proposed from the pulse frequency domain information.The proposed algorithm has the measurement accuracy of 500 Hz and can provide some information to help sorting.(3)In terms of analyzing the node association relationship of Ad-hoc network under IEEE 802.11 standard,a node association relationship analysis algorithm based on IEEE802.11 MAC protocol was proposed.In the analysis of the hierarchical structure of Adhoc communication,it was found that such a RTS/CTS mechanism is generally included in the MAC layer,which can help to reverse mining association information from TOE flows and TOA flows.Combined with the timing constraints between pulses provided by the related information of MAC layer and physical layer in the IEEE 802.11 standard commonly used by Ad-hoc,an association model for three communication modes was established from the perspective of probability.The correlation accuracy of four-way handshake mode and frame segmentation mode is more than 95% in the simulation.(4)In the aspect of general Ad-hoc network node association analysis,a node association analysis algorithm based on continuous communication hypothesis was proposed.Based on the RTS/CTS mechanism and the assumption of continuous communication,we can reduce the dependence on the underlying prior information of time series,mine the association information on a larger time scale,and acquire more robust association information from the accumulation of local association information.In the simulation,the accuracy of model association is more than 90%.