Performance Analysis And Research On Optimal Power Allocation For MIMO Systems Based On Compressed Sensing

With the progress of the mobile communication and wireless Internet, thewireless transmission technology has developed rapidly. The research in the wirelesscommunication field has been focusing on the communication system which has thehigh spectrum efficiency and performance. Multiple-input multiple-output (MIMO)system can achieve much higher spectrum efficiency than the traditional single-inputsingle-output (SISO) system by using the channel state information (CSI) to allocatepower reasonably. In order to obtain higher system capacity, the existing powerallocation schemes assumed perfect CSI in the transmitter, or obtained the partial CSIby channel estimation and feedback. CSI is a key parameter which affects the entirewireless system performance. The existing channel estimation and feedback methodsare difficult to get full CSI in the transmitter. Due to the big system overhead, they arenot suitable for the resource-constrained system. The efficient and accurate CSIestimation and feedback scheme can provide quality of services for the wirelessnetwork.Compressed sensing (CS) theory can recover the original signal utilizing veryfew samples. When the CS theory is combined to the CSI acquisition stage (includingchannel estimation and feedback) in the transmitter of MIMO system, it can use littleinformation to recover channel impulse response accurately. Thus, the spectrumefficiency and the CSI acquisition performance can be improved.A power allocation method based on the determinate observation CS theory forMIMO system is proposed and researched. Firstly, the determinate observation CStechnology is introduced to the CSI acquisition for MIMO system, which means thatthe receiver can estimate nearly complete MIMO channel impulse response from veryfew information. Then at the receiver, the estimated channel impulse response isobserved by determinate matrix and feedback to the transmitter, so that the transmittercan get a nearly perfect recovery of the CSI for MIMO system. The simulation resultsshow that the combination of CS and the CSI acquisition can achieve nearly completeCSI in the transmitter for MIMO system. The design of the determinate observationmatrix can reduce the complexity of observation matrix structure and lower thedifficulty of hardware implementation on the premise of guarantee for acquisitionperformance. In order to get the optimal capacity for the MIMO systems, the suitable power allocation scheme is selected according to the CSI. The scheme assigns morepower to "better" sub-channels, and assigns less power (or probably no power) to"worse" sub-channels. The simulation results prove that the proposed solution canobtain high channel capacity, reduce system overhead and lower the difficulty ofhardware implementation.