Adaptive Double Threshold based Spectrum Sensing for Cognitive Radio Networks

Abstract

Abstract

Amplitude quantization method for autonomous threshold estimation in self-reconfigurable cognitive radio systems

“…technique [19], and the first order statistical technique (FOST) [20], Barne's approach [21], and many other methods as in [22][23][24][25][26]. Essentially, these methods depend on critical parameters that are fine-tuned a priori based on some input noise level or some theoretical distribution of the input sample.…”

Low Power Wide Area Network, Cognitive Radio and the Internet of Things: Potentials for Integration

“…For example, in low SNR conditions, the threshold value of the ED drops too close to the noise floor resulting in increased noise samples crossing the threshold value, leading to increased false alarm rates. There are several options to address this challenge, for example, using effective adaptive threshold techniques such as the forward consecutive mean excision (FCME) algorithm [125,126], double threshold methods [110,117,118,[127][128][129][130], Otsu [114,124], recursive methods [124,131], and several other methods [132][133][134][135][136][137]. These methods would compute new threshold values without prior knowledge of the noise floor in order to prevent excessive false alarms, and to maximize the detection rate of CR-LPWANs.…”

“…The use of adaptive threshold estimation algorithms (ATAs) is advocated in the OOK demodulator block of the SX1276. There are several ATAs available for use such as the FCME algorithm [143], the recursive onesided hypothesis testing (ROHT) algorithm [124], the first order statistical technique (FOST) [144] also called the m-dB method [145], and several other ATAs [128,132,146]. Nevertheless, one research issue to be addressed is the reduction in the computational complexities of these ATAs in order to reduce the power consumption rate of the SX1276 chipset.…”

Generalized self‐tuning system for adaptive threshold estimators in cognitive radio systems using swarm and evolutionary‐based approaches

“…The maximum normal fit method proposed by Barnes et al, also requires the manual fine‐tuning of a few parameters, such as the initial SD and amplitude of the estimated noise distribution 9 . Most methods identified in References 10‐15 typically depend on different parameter values that must be fine‐tuned in order to enhance the performance of the different ATEs. Primarily, most authors often refine the parameters of their respective ATEs by testing them using predefined datasets, or in some other cases, by theoretical computations based on presumed distributions 16 .…”

“…Furthermore, manual tuning can be a cumbersome process, particularly for ATEs characterized by large number of parameters. It is also impossible to obtain global parameter values for every ATE across all possible sensing conditions and signal types, a fact that is evident across most ATEs 10‐15,17,18 …”