Research On Blind Detection Algorithm In Wireless Sensor Networks

The wireless sensor network(WSN)is composed of a large number of small intelligent sensor nodes within the target monitor range.It has the characteristics of low-power consumption,small volume,low cost,distribution and self-organization.Because the energy carried by the sensor nodes in the network is limited,and in most cases the nodes cannot be charged or replaced,the low-power consumption becomes one of the design principles of wireless sensor network.This principle makes the blind equalization and blind detection technology without training sequences become one of the research directions of WSN signal detection technology.On the basis of preliminary results of the project team,the main innovation work of this thesis is listed as follows:(1)On the basis of the study of classical linear prediction algorithm based on the second-order statistics and its improved algorithm,combined with the recursive least squares algorithm,an RLS-MSPA algorithm is proposed.This algorithm not only avoids the inversion solution of output correlation matrix of the linear prediction algorithm,but also show that it has desirable blind detection performance compared with the linear prediction algorithm by experiment.(2)One of the shortcomings of the WSN blind detection system constructed by the Existing literature is that the speed is too slow.The blind detection speed of cluster sensor nodes is crucial to the whole system.In order to improve the system speed,combined with the proposed RLS-MSPA algorithm,this paper constructs a new blind detection model of the RLS-MSPA virtual MIMO wireless sensor network.The RLS-MSPA algorithm is applied to the reference sensor node and the cluster sink node in the cluster.The other non-reference sensor nodes in the cluster employ the related algorithm to restore the original signal by the two-layer signal blind detection inside and outside the cluster to restore the data sent from each WSN node.(3)Because the RLS-MSPA is a blind detection algorithm based on the second-order statistics,it is not suitable for the channel with common zero points.In order to further adapt the system to channels and the system's blind detection performance,this paper introduces an improved ant colony algorithm SSAV-QACO which is applied to the blind detection of reference sensor node signal,and thus proposes an improved ant colony algorithm SSAV-QACO virtual MIMO WSN blind detection system.Compared with the RLS-MSPA algorithm adopted by the reference sensor node,the SSAV-QACO algorithm can obtain desirable blind detection performance and be available to transmission channels,but the time and space complexity are much higher than the RLS-MSPA algorithm's.(4)In order to balance the channel adaptability,complexity and performance of the system,a DS-NSCNN hyper-chaotic precoding virtual MIMO WSN blind detection system is proposed in combination with the hyper-chaotic precoding.Compared with the RLS-MSPA virtual MIMO WSN blind detection system,this system is able to adapt to the majority of channels.Compared with the SSAV-QACO virtual MIMO WSN blind detection system,the proposed system has faster convergence speed and lower complexity.In addition,it outperforms the other two.