MIMO uses 2 to 4 antennas at the transmitter and receiver, which requires individual RF units for each antenna. Massive MIMO uses further large number of antennas (at least 64 antennas) to improve throughput and spectral efficiency, so it is one of the important technology in 5G. Beamforming is a signal processing technique which sends the signal in a particular direction rather than in a broadcast manner, resulting into increased dirctivity, energy efficiency, system gain and signal quality. Hence, it is called as one of the key technology for 5G and adopted by massive MIMO, but it can support only single data stream with one user. Precoding is a generalization of beamforming to support multiple data streams which superposes multiple beams. There are 2 types of precoding i.e. analog and digital precoding, whose combination of these 2 gives hybrid precoding which gives better performance with less complexity. In Successive Interference Cancelation (SIC) based hybrid precoding method, each antenna array is optimized separately (instead of jointly) such that while optimizing the specific array the contribution of optimized array is removed from the total capacity.
These codes are written in MATLAB and are simulated for SIC (Successive Interference Cancelation) based hybrid precoding using mmWave channel in 3D scenario where both azimuth and elevation angles are taken into account and antennas at the Tx and Rx are arranged in 2D i.e. UPA or URA, in other words full-dimentional massive MIMO is considered here. These codes are simulated for multi-user and multi-cell cases and also compared their performance with different parameters. Also codes compares different factorization methods (like SVD, EVD, GMD) for capacity optimization in hybrid precoding.
Below described the functions/operations performed by different files -
- SIC_based_HP_3D_MU_FC_SC.m - It simulates SIC based hybrid precoding and optimal precoding schemes using sub connected and fully connected structures separately for multi-user case in 3D scenario.
- SIC_based_HP_3D_MU_FC_SC_comparisions.m - In this code the achievable rate of different schemes i.e. SIC based hybrid precoding and optimal precoding schemes using sub connected and fully connected structures separately for multi-user case in 3D scenario, are compared with the different parameters like number of Tx/Rx antennas, number of RF chains at Tx/Rx, number of paths/rays and with the number of users. Here to compare with a specific parameter, you have uncomment that specific section for that parameter in the code (comment all other sections related to other parameters) and run it.
- SIC_based_HP_multicell.m - Simulates SIC based hybrid precoding and optimal precoding schemes using sub connected and fully connected structures separately for multi-cell case in 3D scenario.
- SIC_based_HP_multicell_comparisions.m - Here the achievable rate of different schemes i.e. SIC based hybrid precoding and optimal precoding schemes using sub connected and fully connected structures separately for multi-cell case in 3D scenario, are compared with the different parameters like number of Tx/Rx antennas, number of RF chains at Tx/Rx, number of paths/rays, number of users per cell and with the number of cells. Here to compare with a specific parameter, you have uncomment that specific section for that parameter in the code (comment all other sections related to other parameters) and run it.
- SIC_based_HP_3D_MU_FC_SC_using_EVD.m - Compares between SVD and EVD for capacity optimization for multi-user case in 3D scenario, where capacity optimization is carried on the SIC based and optimal precoding algorithms using sub connected and fully connected structures separately.
- test_2.m - It is a simulation code for GMD based hybrid precoding using mmWave channel in 2D scenario. which uses OMP algorithm to find analog precoder and uses GMD for calculating digital precoder. Acually this is not a complete code, whose work is currently in progress (Code will be updated when simulation completes).
Note that, in 2D scenario antennas will be arranged in linear i.e. ULA and only azimuth angle is considered in the channel.