The Research On Interference Alignment With Finite Extensions In Wireless Networks

With the blowout growth of data flow,in order to provide high quality communication network services,5G is committed to changing the communication architecture and providing new communication modes.It can provide higher peak rate,lower delay to millisecond level,and great connection density.However,this communication architecture will cause serious interference problems.In a communication system comprised of multiple cells and multiple users,interference is an important factor hindering the communication system performance.In order to eliminate interference,various interference management strategies have become a research hotspot in this field.Among them,interference alignment(IA),as a novel technology,has attracted intensive research interests and has become a promising project.This paper investigates the feasibility of IA for multi-input-multi-output(MIMO)interference broadcast channel(IBC).Traditional IA is mostly investigated in the scenario of infinite channel extensions or no channel extensions.Infinite channel extensions are not feasible in practical communication systems,and the performance of no channel extensions can be further improved.Therefore,this paper will study IA in the scenario of finite channel extensions.In this paper,the problem of feasibility of IA in common message MIMO interference broadcast channel(IBC)wireless networks in three different scenarios of channel extensions have been investigated.Moreover,corresponding necessary conditions have been derived which provide constraints that network configuration parameters must satisfy.These constraints restrain the maximum achievable degree of freedom of the network.First,in the scenario of no channel extensions,the channel matrix is structureless.Based on algebraic geometry,this paper derives a necessary condition by establishing the equivalent relationship between the feasibility of IA and the solvability of multivariate polynomial equations.Second,in the scenario of T finite independent channel extensions,the channel matrix is block diagonal.This paper resolves the total system into two subsystems and derives necessary conditions for these two subsystems to obtain a necessary condition for the totalsystem.Thirdly,in the scenario of T finite dependent channel extensions,the channel matrix is not only block diagonal but also channel coefficients are no longer generic.This paper derives a necessary condition by transforming the problem of feasibility of IA with nongeneric coefficients to the problem of feasibility of IA with generic coefficients by applying channel diversity.In this paper,the performance limit of the network degree of freedom is demonstrated by numerical results,which provides theoretical guidance for the design of interference alignment algorithm.