Reseach On Key Technologies Of Autonomous Cognition For Space Communication Network

Space communications has a worldwide prospect based on autonomous cognition technologies. It helps to sense, learn and reason for the network itself; it also enhances the decision-making capabilities via artificial intelligence technology, and even contributes to fast realizing flexible connection between the heterogeneous subnetworks and their nodes.On the background of distributed payloads and wireless network detection technology, based on autonomous cognition technologies, this thesis focuses on the issues of the space communications network structure design, the inter-orbit channel modeling method, as well as the signal modulation recognition scheme:First of all, based on autonomous cognition technology, a space communications network structure with heterogeneously fusion is proposed. This structure is inspired by the vertebrate nerve system structure and human society organization mode; its corresponding structure for network-level, sub-network group level and node-level have also been designed and explained.Furthermore, the geometry parameters and characteristics’ models of the inter-orbit satellite links has been set up. Then, the space channel model has been mainly divided into three modules, a free space loss module, a Doppler shift module and a noise module; afterwards, the regulations between channel models and the inter-orbit parameters have been explicitly pointed out by Matlab modeling and simulation.Last but not the least, a signal modulation recognition scheme with a hybrid classifier is designed to be applicable to non-cooperative signals. Six classes of parameters are designed to reduce the sensitivity to noises; the corresponding judgment thresholds have been confirmed by simulation based on decision theory. Then, the algorithm has been enhanced by combining an improved BP neural network algorithm and the decision tree above. Simulation results show that: when SNR is no less than 0 d B, the average success rate of this scheme for 13 kinds of modulations is above 98%.