Research And Improvement On Weak Signal Detection When Primary User Has Multiple Transmit Power Levels

Cognitive radio is an efficient technology to solve the conflict between the growing congestion and the inefficient use of spectrum bands in wireless communications.Spectrum sensing is the core to realize the function of cognitive radio.Only when the secondary user detects the usage of spectrum correctly,it can make a right choice,namely,use the spectrum or give up it.Common specturm detection methods include energy detection,matched filter detection,etc.But,it is not difficult to notice that nearly all the sequential spectrum sensing schemes assume that primary user either being absent or being present with a constant power level.Yet,we know from many current standards,i.e.,IEEE 802,11 series,LTE,LTE-A that the licensed users could be working under different transmit power levels in order to cope with different situations.Besides,with the possible huge fluctuation of noise and the weak signal received by secondary user,it is hard to sense the specturm correctly.Considering above problems,this paper mainly studies the weak signal detection and recognition problem for this multiple primary transmit power levels scenario.(1)The basic content of cognitive radio is introduced and the spectrum sensing technologies are analysed.Considering that the existing spectrum sensing technologies assume a constant power level of primary user,we propose the fixed detection for this multiple primary transmit levels scenario.The target is not only to detect the presence of primary user but also to tell the transmit power level of primary user in this scenario.Specificly,we propose two detection schemes.The thresholds over existence of inexistence in these two schemes are compared,the performance analysises are also carried out between these two schemes.(2)Considering that fixed detection can not deal with the problem of long sensing time when the signal is weak,which leads to low spectrum utilization.We improve the sensing schemes and propose a sequential spectrum sensing scheme for this multiple primary transmit power levels scenario.The main advantage of the sequential spectrum sensing is that it can reduce the average number of independent samples remarkably to achieve the same detection performance as that of fixed-sample testing methods,i.e.,it can improve the spectral efficiency.Specificly,we propose two different but relevant sequential sensing schemes,i.e.,detection before recognition and recognition before detection.We derive the closed form expressions of decision regions and provide the performance comprison and simulation between these two methods.(3)We systematically discuss all aspects around the sequential spectrum sensing with multiple primary transmit power levels,including the variation of the decision regions,the closed-form average sample number,compution complexity and cause of the power mask phenomenon,performance comprison with fixed detections,etc.Numerous simulations are also carried out to verify the analysises.