| Mobile Ad Hoc Network(MANET)is a communication network without central infrastructures such as traditional base stations.In recent years,MANET has been widely applied on military wireless communication,emergency communication,rescue and relief work,and other emerging applications due to its advantages of flexible topology and fast networking.The National Guideline on Medium and Long Term Program for Science and Technology Development proposes that the Ad-hoc network technology is taken as the frontier development in China.As the reference for the design and evaluation of communication systems,wireless channel models play a pivotal role in future MANET technologies.However,the researches on channel measurement and modeling of urban and air-ground scenes are still inadequate,the statistical channel models of these related scenes are also insufficient.Moreover,because the antennas of MANET nodes are low from the ground,the factors affecting signal propagation are complex.The performance of traditional channel parameters estimation algorithms declines when affected by noise and interference.Therefore,it is urgent to establish a more accurate and robust channel parameter estimation scheme.Aiming at the channel measurement,channel modeling,and channel characteristics extraction algorithm under MANET scenarios,the main work of this paper is as follows:1.The principle and applications of existing channel measurement methods and measurement data post-processing methods are analyzed in every detail.A channel characteristics extraction scheme that satisfies the requirements of channel modeling is summarized.In the scheme,the thresholds based on constant false alarm rates are applied to remove measurement noise.The extraction methods including path loss factor,delay spread,Doppler spectrum,and other feature parameters are given.Furthermore,the channel sounding test platform based on USRP X310 and Propsim F8 channel simulator is established,and the effectiveness of the characteristics extraction algorithm is verified.2.To establish the urban and air-ground wireless channel models,channel measurements in VHF and microwave frequency bands were carried out in Qingdao and Mangshi,China.A large number of channel impulse response data were obtained.Based on the measurement data,the influence of the receiver’s mobility and scatterers’ distribution on channel fading is analyzed.The optimal statistical models for channel characteristics are selected by statistical tools such as Akaike’s information criterion and hypothesis testing.Furthermore,the large-scale fading model and multipath channel model for urban non-light of sight,urban quasi-light of sight,and air-ground communication scenarios are given,which can provide a reference for the design and optimization of future mobile communication systems.3.Motivated by the defect of model overfitting in SAGE(Space-alternating Generalized Expectation-maximization)caused by channel noise,a new parameter estimation scheme based on variational sparse Bayesian learning and Gaussian mixture model is proposed.To depict the complex fading phenomenon in MANET scenarios,the channel coefficients are represented by Gaussian mixture models(GMM).This scheme uses sparse Bayesian learning and variational inference to deduce the posterior probability distribution of channel variables.Benefiting from the flexibility of GMM,the scheme can represent various fading channels while controlling the order of the channel model.Simulation results show that the proposed algorithm has significant advantages over the existing algorithm in terms of estimation accuracy. |
