Research On Spectrum Sharing Technology Under Complex Dynamic Environment

With the advent of the information revolution,spectrum resources have become increasingly paramount in maintaining social stability and eco-nomic prosperity.The continuous and rapid development of the commu-nication industry has remarkably improved the level of social informati-zation,meanwhile exacerbated the scarcity of the spectrum resources.As a promising way to enhance spectrum efficiency and alleviate spectrum shortage,spectrum sharing has attracted extensive attentions.On the oth-er hand,in order to support diverse data transmission requirements,the dense heterogeneous and dynamic characteristics of further wireless net-work would be extremely prominent.Those complex features of wireless environment will pose a more severe challenge to spectrum sharing.To cope with the dense,heterogeneous and dynamic characteristics of wireless communication network,this dissertation focuses on the two as-pects of spectrum sharing,i.e.efficient utilization and reliable access.We specifically investigate the spectrum sharing issues in the ultra dense net-work,the dynamic heterogeneous network and the highly dynamic net-work.Accordingly,the main contributions are listed as follows.1.Due to the low utilization efficiency of the most existing tempo-ral domain spectrum sharing schemes,how to achieve a high spectrum efficiency to support tremendous access demands of dense wireless user-s is a major obstacle.To address this challenge,we take the ultra dense network as an example,and design a space reuse based two-dimensional spectrum sharing scheme,where multiple users are allowed to access the same channel simultaneously.Given the selfish and rationality of users,a non-cooperative game with fine-grained two-dimensional reuse is formu-lated.It is then proved to be an ordinary potential game,which guarantees the existence of the strategy Nash equilibrium.To reduce the signaling overhead for global information exchange,a decentralized reinforcement learning algorithm is proposed,which relies on the individual information to update the strategies.The convergence efficiency of the new scheme is rigorously proved.Numerical simulations are provided to validate the performances of the proposed scheme.2.Due to the strict dependence of traditional spectrum sharing meth-ods on ideal channel state information,how to guarantee a robust spec-trum access in the practical dynamic wireless environment without accurate channel information is a key problem.To solve this,we take the dynam-ic heterogeneous network as an example,and propose a channel statisti-cal information based spectrum sharing framework,which eliminates the dependence on complete channel information.On this basis,for the spec-trum reuse,we adopt a correlated equilibrium-based game to formulate spectrum reuse,and propose a distributed regret-matching learning algo-rithm to achieve the solutions.With the introduced "regret",the proposed scheme can obtain more promising solutions.For the joint power and chan-nel optimization,by decoupling it into two subproblems,a joint resource allocation algorithm is proposed,which relies only on the statistical infor-mation of dynamic channels to implement.Numerical results are provided to corroborate the anticipated performances of our proposed schemes.3.Further considering the emerging highly dynamic network scenar-ios,the primary challenge of spectrum sharing is how to achieve reliable and effective spectrum access in short competition slots.To this end,with the fuzzy theory introduction,we establish a fuzzy space based spectrum sharing scheme,and take the unmanned aerial vehicle(UAV)network and the vehicle-to-everything(V2X)network as examples.Specifically,for the partially overlapping channels assignment problem in a mesh UAV net-work,we adopt a fuzzy payoffs game to formulate this problem,and pro-pose a robust and distributed fuzzy learning scheme to achieve the fuzzy Nash equilibrium.For the joint time-frequency allocation problem in a cel-lular V2X network,we construct a two-side many-to-many fuzzy match-ing game(MM-FMG)to formulate the optimization problem,and propose a dynamic vehicle-resource matching algorithm to solve this MM-FMG problem.Comprehensive simulation results demonstrate that the designed fuzzy space based learning schemes are superior to the conventional meth-ods,and can achieve robust channel access in the highly dynamic and com-plex wireless environment.Finally,the investigations of the whole dissertation are summarized.On this basis,several interesting research directions on spectrum sharing technology in future communication networks are discussed.