Primary User's Activity Prediction Based Spectrum Sensing And Path Selection Scheme

Cognitive Radio(CR)technology has been proposed as a key technology to resolve the spectrum crisis of the ISM band.The operation of secondary users(SU)is affected by the primary user(PU).On the one hand,in order to avoid interference with the PU,the SU needs to perform frequent spectrum sensing on the licensed channels for effectively monitoring primary user(PU)activities.However,frequent spectrum sensing will greatly increase energy consumption and reduce the life-time of the networks.On the other hand,the transmission stability of secondary users is affected by the primary user.When the PU activity is frequent,the secondary user has to frequently interrupts transmission to avoid interference with the primary user,resulting in unstable link and re-transmissions,thereby increasing energy consumption and reducing the throughput.Therefore,reducing the negative impact brought by the PU,decreasing the energy consumed in spectrum sensing,and ensuring the stability of the communication link are particularly important for cognitive radio sensor networks(CRSN),which has limited resources.This thesis focus on the PU activity prediction based optimizing scheme,the main work and innovations are as follows:In order to optimize the energy efficiency of spectrum sensing and eliminate redundant nodes in the traditional scheme,a Hidden Markov Model(HMM)based redundant node elimination scheme is proposed.Meanwhile,a representative nodes selection scheme based on the global sensing accuracy and residual energy is proposed to optimize energy efficiency in cooperative spectrum sensing.Experimental results show that by eliminating redundant nodes and representative node selection,the scheme can effectively reduce the energy consumption of spectrum sensing and improve the network energy efficiency.Due to the difficulty in collecting labeled data in multi-hop cognitive radio sensor networks,a primary user activity prediction model based on semi-supervised learning is proposed,which optimizes the number of labeled samples.In order to reduce the impact of primary user activities on secondary users,a new path selection standard is defined,which is defined as the reliability of communication.This new standard reflects the probability that relay nodes can conduct successful transmission,thus reducing the impact of primary user activities on secondary users.Experimental results show that the proposed scheme can effectively reduce the storage cost of cognitive radio sensors and guarantee the transmission stability.