Modulation Classification In Successive Relaying Systems

The successive relaying system is a new method to mitigate the 50% throughput loss in traditional relay systems.In the successive relaying system,the relay node adopting decode-and-forward(DF)strategy can avoid system noise amplifying in amplify-and-forward(AF)strategy.In the DF strategy,we need to know the modulation type of signal.However,when the successive relaying system applied in the military or the cognitive network,the communicating nodes are non-cooperative,so the modulation type of the signal is unknown.Therefore,we need to automatically classify the modulation of the source signal.Aiming at classifying the source modulation in successive systems,this paper conduct research as follows.Since no previous work address the problem of classifying modulation with interference in multi-node communication systems.We automatically classify the modulation with inter-relay interference in successive relay systems based on the likelihood function,since the likelihood-based method is optimal in Bayesian sense and can provide upper bound for other methods.Firstly,the likelihood function of the received signal on some modulation pair of the source and interference signal is derived.And it can be seen as the conditional probability function under some modulation pair of the source and interference signal.After that,the edge probability relating to the source modulation was derived.As a result,we can decide that the source modulation type which makes the likelihood function get the largest value is the classified modulation type.Since when we derived the likelihood function,the interference signal is considered as a variable,the performance of our proposed algorithm can be improved.And we prove it both in theory and simulation.To reduce the computational complexity and make our algorithm of jointly classifying interference more practical,we also classify the modulation according to the higher-order statistics(HOSs),which is called feature-based method.We extract the HOSs of the received signal as classifying features.And under the assumption that they obey the Gaussian distribution,we compute the mean values and the variance matrix of HOSs taking into account the interference signal.So the probabilistic distribution of the features is related to both the source and the interference signal.Then according to the joint probability,we derive the edge probability relating to source modulation which is utilized to construct the likelihood ratio classifier.Simulations show that the proposed algorithm can efficiently suppress the influence of the interference signal.Lastly,based on the fact that in the actual successive relaying communication systems,the interference signal may not be synchronized,we take the timing offset of the interference into account.And we derived the probabilistic distribution of the features related to the timing offset,after that we get the maximum likelihood estimator of the timing offset.Then we construct the likelihood ratio classifier under the condition of the estimated timing offset.Simulations show that the modified algorithm can resist the timing offset effect.