Multicell Coordinated Interference Suppression Technology In Wireless Cellular Systems

To accommodate the rapidly growing number of users and the exponentially increasing demand for data rates,universal frequency reuse is adopted to enlarge the cell coverage and improve the system performance in future wireless networks.However,using the same frequency for adjacent cells can cause severe interference,especially for the cell-edge users whose quality of service is seriously declined.Therefore,how to suppress the intercell interference is an urgent and critical problem in wireless cellular systems.As one of the most concerned technologies,multicell cooperation allows the base stations to share the channel state information(CSI)and the user data,and then jointly design the transmission scheme to deal with the interference.This thesis deeply studies some important problems in multicell coordinated interference suppression technology.The transceiver and the resource allocation are carefully designed by considering the limiting factors in actual systems.Firstly,we study the weighted sum-rate maximization problem and the global optimal precoding design in multicell multiuser multiple-input single-output systems.By reformulating it into a monotonic optimization problem over the achievable rate region,a novel monotonic optimization algorithm is proposed to iteratively approach the optimality.The proposed algorithm maintains a set of non-overlapping hyperrectangles that contains parts of the achievable rate region where the optimal solutions can be potentially obtained.By iteratively dividing,reducing and removing the hyperrectangles,the range of the potential region is gradually narrowed.Each iteration includes a sensible search scheme,a sequential partition method and a vertex relocating procedure.Comparing with the existing methods based on the achievable rate region,the proposed algorithm significantly reduces the computational complexity and accelerates the convergence.Secondly,the transceiver design for multiple-input multiple-output(MIMO)interference channels is investigated when the CSI has statistical error.The users can transmit multiple streams without violating the feasibility of degrees of freedom.In terms of the bit error rate(BER),we choose the maximum stream mean square error(MSE)as the objective.The statistical CSI error model,which describes both the inaccurate estimation and the CSI delay,is stated.We calculate the average stream MSE and then formulate the maximum average stream MSE minimization problem.The alternate optimization approach is adopted to iteratively update the transmitter and the receiver.When the transmitter is given,we can obtain a closed-form receiver.As to the transmitter update,a centralized algorithm based on the second-order cone programming(SOCP)and a distributed algorithm adopting Lagrangian duality are proposed.The convergence and computational complexity are analyzed.The centralized algorithm is proved to converge to the set of Karush-Kuhn-Tucker solutions while the distributed approach has much lower complexity.Moreover,the distributed implementation and the communication overhead in both time division duplexing and frequency division duplexing systems are described.The simulation results validate the effectiveness of the proposed algorithms on dealing with the statistical CSI error and the improved BER performance over the existing methods.In addition,it reveals that the distributed algorithm can significantly reduce the computational time without sacrificing the BER performance.Thirdly,we consider another CSI error model for the transceiver design in MIMO interference channel.The CSI error is randomly distributed within a bounded area.A slack variable is introduced to represent the worst-case stream MSE,and then a problem minimizing the maximum worst-case stream MSE is formulated.Since the error has infinite realizations,the number of constraints in the problem is also infinite.Two methods are proposed to tackle it.One acquires an equivalent problem using S-procedure while the other one obtains an approximated problem via the properties of inequality and norm.The alternate optimization approach is utilized for both problems.The simulation results show solving the approximated problem has lower complexity,thus is preferable for systems with a large number of antennas.Compared with the methods focusing on the total MSE and the user MSE,the stream MSE based algorithms can achieve lower BER.Besides,the robustness to bounded CSI error is also validated in simulation.Lastly,taking the practically rate-limited data-sharing backhaul link into account,we investigate the coordinated transmission scheme,including the base station clustering and precoding design in multicell heterogeneous networks.The backhaul network is in a tree structure.The central unit is the root while the base stations form the nodes in the tree.Due to the fact that each node has only one incoming link,the relation between the backhaul rates and the wireless transmission is formulated as linear inequalities of binary variables.A problem that simultaneously optimizes the base station clustering and the beamforming is formulated to maximize the minimum rate among users.For this non-convex mixed integer optimization problem,we proposes two low-complexity suboptimal algorithms based on linear programming relaxation.The first one conducts the bisection method for the objective and proposes a three-step algorithm for the feasibility problem.The three steps are power allocation,base station clustering and feasibility checking.The second algorithm also contains three steps: the minimum rate optimization without backhaul capacity constraints,base station clustering,and the minimum rate optimization with given clustering scheme.We define an importance factor for each wireless link and then formulate an importance maximization problem to determine the base station clustering.This integer linear programming is solved via its relaxation.The optimality condition is also discussed.The simulation results manifest that the proposed algorithms can achieve near-optimal solution and guarantee the backhaul capacity constraints.