Signal Processing for Sensor Networks With Compressive Sampling

Abstract

ABSTRACT newlineSensor Networks (SNs) are becoming an inevitable part of the processes in almost all the newlinefields. Networking a large number of varied sensors to monitor, collect, disseminate and newlineintelligently combine information about a specific task is the emerging paradigm of the era. newlineThe flexibility, fault tolerance, high sensing fidelity, low cost and rapid deployment characteristics newlineof SNs create many new and exciting application areas for them. As the usage and newlineapplication areas are becoming wider, the design and implementation of SNs face variety newlineof challenges. The conventional signal processing methods in SNs are based on the Nyquist newlineSampling theorem, which requires sampling rate at least twice the message signal bandwidth newlinein order to achieve exact recovery. This gigantic amount of data received from these sensors newlinemust be processed to estimate, detect and predict various phenomena, which needs huge newlinebandwidth, processing power, storage space etc. The huge data generated by these systems newlineare creating overload for the systems of processing, communicating, storage, power and control. newlineThis area needs a lot of contributions to cover up the issues to have properly oriented newlineand well functioning networks. newlineCompressive Sensing (CS) is of interest in the scenario where sampling by Nyquist newlinerate is not efficient. CS can find sparse solutions to underdetermined linear systems and can newlinereconstruct the signals from far fewer samples than suggested by the Nyquist rate. The research newlinefocuses on the field ofWireless Sensor Networks (WSNs). After a detailed study and newlineimplementation of a typical WSN, it investigates the signal processing activities and challenges newlineinvolved in WSN. The research tries to elucidate the drawbacks of traditional signal newlineprocessing methods. The data deluge problem generated by the SNs has been portrayed to newlineunderstand the critical situation pertaining in the field. newlineThis research is an attempt to redefine the WSNs with the new concept of CS, in which newlinethe sparsity of the signals can be exploited to avoid the wastage of resources. The traditional newlinemethods are modified and the regeneration of sensor data with lesser samples than Nyquist newlinerate has been accomplished with the CS methods. Along with this, the potential of CS to newlinesolve the issues mentioned for the WSNs that aroused as a result of data deluge due to high newlinesampling rate, is examined. The capability of CS has been proved with different performance newlinemeasures and concluded that CS can effectively solve the existing issues in SNs. newlineiii newline