Research On Joint Allocation Algorithm Of Spectrum Power Based On Game Theory In Underwater Acoustic Communication Networks

A lot of countries have turned their attention to marine development because of the serious resources shortage on land.However,the lack of effective underwater communication methods hamper human's footstep in developing ocean seriously.The underwater communication environment is complicated.The wireless communication technology,which has developed maturely on land,is difficult to be employed underwater directly.The sound has been proven to perform better than electromagnetic and light underwater.However,underwater acoustic communication faces many difficulties: severe attenuation of underwater acoustic signals,limited available spectral resources,and limited energy supply.The underwater spectrum and power resources are valuable.Reasonably allocating spectrum and power resources to improve the performance of underwater communication networks is the main issue of this study.Firstly,the characteristics of underwater acoustic communication channels are analyzed,including: underwater channel loss,underwater frequency characteristics,underwater acoustic signal propagation velocity,underwater Doppler effect,etc.And the underwater acoustic channel characteristic is compared with the wireless communication channels on land.At the same time,the basic concepts of game theory,the classification of game theory and the model of stochastic game are introduced.Secondly,the problem of power allocation in underwater single-channel communication networks is studied.The power allocation problem in the single channel is constructed into a game model.As for mutual interference of nodes in the single-channel network,the network interference level is introduced into the effect function to balance the interference level among nodes.For the problem of energy limited,the user residual energy is introduced into the effect function and improves the user's survival time effectively.Finally,the problem of joint allocation of spectrum and power in underwater multi-channel communication networks is studied.An autonomous learning algorithm for joint spectrum and power allocation under dynamic underwater communication environment is proposed.We define the joint spectrum and power level allocation problem as a stochastic non-cooperative game.The process of selecting spectrum and power is constructed as a Markov decision process.In the game,each user maximizes their long-term reward by learning local information without information exchange among users.And the Q-learning algorithm is used to learn its optimal strategy in a distributed manner.As for the complex and changeable underwater communication environment,the supervisory factor is introduced into the Q-learning algorithm to enhance the anti-jamming capability of the algorithm and improve the success information transmission rate of users.