A Study Of User Scheduling In Downlink LTE-A System

CoMP(Coordinated Multiple Points Transmission/Reception), as one of the key technologies in LTE-A(Long Term Evolution-Advance)system, has become the research hotspot in recent years. Co MP which makes fully use of the coordinated transmission among multi-cells and multi-users, efficiently suppresses interference between edge users, improves service quality of edge users, and eventually improves system performance. Many edge users remain to be served. However, the number of users that can be served in the same time-frequency resource is limited. It is necessary to select a set of suitable users with the help of scheduling algorithms to achieve better system performance. Taking implementation into consideration, the complexity of scheduling algorithms can’t be too high.The channel capacity of MIMO(Multi-input Multi-output) system and precoding technology which eliminates interference between users are studied in this paper. For block diagonalization precoding, some algorithms which are improved in the mathematical method are introduced. The improved block diagonalization precoding methods reduce the computation of matrix factorization with the performance of system unchanged. Then the paper focuses on downlink multi-user scheduling algorithms and analyzes some algorithms with low complexity. At last the improved methods in two ways are proposed in this paper. The improved algorithms in this paper still apply greedy thought. Every time one user is selected into the selected set of users and the number of the remaining set of users is reduced gradually. When the number of the selected users reaches limit or the sum capacity begins to decrease, the scheduling process terminates. Compared to algorithms based on capacity, the improved algorithms change scheduling guideline and overcome the high complexity problem caused by singular value decomposition. The guidelines in this paper consider some aspects below. One aspect is to assume that users can cooperate with each other. Then the single user capacity can be thought of as an upper bound of multi-users capacity. According to the assumption, the user in each step that can provide the highest bound with those selected users is found. Another aspect is to consider the correlation between users which diagonalization precoding is sensitive to. The user in each step that has the lowest correlation with those selected users is found. The improved algorithms eliminate a large computation caused by matrix factorization. By simulation results, the improved algorithms reduce the complexity with small losses of sum throughput.The classic greedy algorithms are applied to solve joint scheduling of users and antennas in this paper. In joint scheduling process, every user does not use all its antennas to receive signals. Because of this, more users can be scheduled at one time, thereby achieving a higher sum capacity. As the number of the antennas increase, the correlation between antennas becomes the key factor which impacts the total throughput. Two low complexity joint scheduling algorithms based on correlation are introduced in this paper. Firstly the angle between antennas is used to measure the correlation. Secondly, the Chordal distance is used to measure the correlation. The simulation result shows the introduced two algorithms achieves comparable total throughput as greedy algorithm with lower implementation complexity.