Blind Detection System Based On Clustering Virtual MIMO Wireless Sensor Networks

Wireless sensor networks (WSN) is the network which a large number of sensors deployed in the target area are connected by the way of self-organization. Sensors are battery-powered in most cases, and the energy is limited, and the design of energy-saving of wireless sensor networks become one of key issues, the equalization of current wireless sensor networks all use adaptive equalization techniques based on training sequence, however, because of blind equalization technique dose not use training sequence and it can improve transmission rate of effective information, thus it can saving energy for the sensor. So it is valuable to research for blind detection based on wireless sensor networks. The main work of this paper is as follows:(l)We study and analyze the blind equalization principle and the classification of blind equalization algorithms. This paper focus on blind detection algorithm based on Hopfield neural network, and points out due to the algorithm need small(tens of) data and there is no restrict whether the public zero channel is exist or not when processing data, it can be applied to wireless sensor networks.(2) We firstly study and analyze the transmission policy and related technique raised in some literature about researching on energy-saving of wireless sensor networks. Which points out the transmission model need training sequence when transferring data whether it is the planar structure or layered structure, however, blind detection technique does not require training sequence, which can save energy for sensors. Then the paper combines the feature of high-density of sensors deployed in wireless sensor networks and virtual MIMO (Multi-input Multi-output) technology, constructs a blind detection system based on clustering virtual MIMO wireless sensor networks.(3)Against the internal model of blind detection system this paper constructs, we combine the theory of correlation and blind algorithm based on Hopfield neural network, propose a blind scheme and also analyze the feasibility of the scheme. The scheme first use blind detection algorithm based on Hopfield neural network to derive the reference node information, and then combine with the correlation among the transmission signal of each sensor of the model which the paper constructs to recovery other nodes information in the cluster. Simulation results show the feasibility of the scheme.(4) Against the external virtual MIMO model of blind detection system this paper constructs, by introducing signal space cancellation (SSC) and successive interference cancellation (SIC), the blind detection algorithm based on Hopfield neural network is extended to virtual MIMO system, finally we verify the feasibility of this algorithm through simulation experiments.