No abstract

No abstract

A newly deployed sensor network is unstructured and sensor nodes at the beginning are in the "blind" state, which have almost no knowledge about the topology of the network. The most urgent and important thing for deployment of sensor networks is to set up a coarse structure. In this paper, a practical algorithm to initialize the network and select cluster leader nodes is presented. The algorithm is analyzed on SNIR model, which addressed the specific characteristics in the deployment of wireless sensor network, such as nodes do not feature a reliable collision detection mechanism and may wake up asynchronously, and especially co-channel interference effect on transmission during initialization phase. A novel concept of interference-tolerant neighbor nodes is proposed to guarantee efficiency of the algorithm. The simulation result shows even in the restrict applicable model, clustering can be completed successfully.

To classify the frequency modulation signal, this paper employs a parameter invariant •ter, which can transfer the frequency modulated information to variety of its envelope, and then extracts the histogram f~ to classify the modulation type. This method can efficiently classify the type of a signal such as frequency mo~dafion (FM), binary frequency shi~ keyli~ (BFSK), quadrature frequency shift keying (QFSK), 8-ary frequency shift keying (8FSK), etc. It can easily be realized and is especially suitable to applications in soft, rare radio.

No abstract