Performance Analysis Of Interference Alignment Algorithm Based On Antenna Selection And Its Hardware Design

Wireless access can already be implemented on a global scale, making wireless spectrum resources strained, so how to improve the spectrum utilization efficiency in wireless communication system under the condition of high traffic requirements has become a key study topic. With the exploding users, interference becomes the bottle neck problem to improve internet capacity. To solve this problem, interference alignment (IA) is proposed. As a novel processing method of interference, the main idea of IA algorithm is aligning the interference signal to the same interference subspace at the receivers, so there are signal dimensions free of interference at the receivers. Two iterative IA algorithms-distributed MinIL and Max-SINR algorithm are representative and are studied widely.In IA algorithm, the users in the system maybe in sleeping mode, so there are redundant antennas. Since the cost of RF chain is higher than antennas, so choosing a proper set of antennas at receivers or transmitters is very meaningful. In this paper antenna selection (AS) is applied into MinlL and Max-SINR algorithms and studied in different conditions. Besides, the way to run hardware simulation of IA algorithm based on Quartus Ⅱ and Modelsim is introduced. The work accomplished in this paper is shown as follows:(1) Simulate and analyze two classic IA algorithms-MinIL and Max-SINR algorithm. The results show that when SNR is low or medium, Max-SINR can achieve a higher sum-rate than MinIL, and the outage probability of Max-SINR algorithm is lower than MinIL as well. When the SNR is high, the performance of two algorithms are the same.(2) Combine AS with MinIL and Max-SINR algorithm and run the simulation, and prove that AS can improve the sum-rate of system. The results show that when users in the system are less than(M+N)/d-1, if receivers use the same number of antennas to receive signals, system applied with distributed MinIL and Max-SINR algorithm combined with AS can achieve a higher sum-rate. But when receivers use all antennas to receive signals, if the SNR is low or medium, Max-SINR algorithm that uses all the antennas can achieve a better performance than AS algorithm. When the SNR is high, Max-SINR applied with AS can achieve a higher sum-rate than Max-SINR that uses all the antennas. Besides, MinIL algorithm applied with AS can always achieve a better performance than MinIL algorithm that uses all the antennas regardless of SNR. Then this paper proves that when the number of users are less than (M+N)/d-1, the distributed Max-SINR algorithm can improve the Max-SINR of remaining users. The computation complexity of distributed MinIL algorithm, distributed Max-SINR algorithm and when the two algorithms are combined with antenna selection algorithm is analyzed as well.(3) Implement the hardware design and simulation of IA algorithm through Quartus Ⅱ and Modelsim. Discuss two methods of getting the minimum eigenvalue and its corresponding eigenvector of matrix-definition method and inverse power method. Besides, this paper also designs entities to extract the minimum eigenvalue and its corresponding eigenvector, add matrix, multiply matrix and so on. Then implements MinIL algorithm in a MIMO system and tests it through hardware simulation. The result shows that interference is eliminated effectively. Finally, the key step of Max-SINR algorithm hardware design-getting the inverse of matrix is implemented.