| The wireless sensor network is a new concept proposed in recent years. It is a newkind of distributed information access system and is a combination of moderninformation processing technology. Its emergence has caused widespread concern in theworld. It’s regarded as one of the most important technologies that will have asignificant impact on the21st century.Wireless sensor network monitoring electromagnetic spectrum of communicationenvironment is an important aspect of its many applications, including signal detection,identification, listening and positioning. The identification and classification of thesignal modulation is between signal detection and signal demodulation, thus is a keystep to listen signal. How to acquire a high accuracy of signal identification of sevendigital modulations at a low SNR is the main content of this paper.Based on the identification requirements of digital modulation of wireless sensornetwork, we analyzed the research status at home and abroad in this area. After that wedesigned an identification method by using high order cumulants and support vectormachine and introduced each module of our algorithm. Then we applied our algorithmon the project’s hardware platform from top level to bottom level to form a completeidentification functional module. During the hardware implementation of our algorithm,we used several appropriate technologies to solve the various problems we encounteredin the practical situation. In addition, most of the identification algorithms weredesigned and simulated for one single node, while in our paper we used the method ofdistributed collaborative identification and also implemented it on our hardwareplatform.After algorithm design and implementation, we did a lot of tests by using actualsignal. The results indicated that our algorithm was able to identify seven kinds ofdigital modulations with high accuracy at a lower SNR and that distributedcollaborative identification did improve and enhance the performance of one singlenode. |
PDF Full Text Request