Study Of Multicell Coordinated Precoding

The next generation cellular system adopts Orthogonal Frequency Division Multiple Access (OFDMA) in the downlink (DL) channel to improve spectrum efficiency with full frequency reuse. The main interference is from adjacent cell, viz. inter-cell interference (ICI), and intra-cell interference can be ignored. To further improve system throughput and the cell-edge users’ performance, reducing ICI is becoming a very important research topic. Coordinated Multi-Point Transmission/Reception (CoMP) is one of the key technologies of the4-th generation (4G) cellular mobile communication system. In the DL, by sharing information across BSs and designing downlink signals cooperatively, the traditional interference can be converted into useful signal, and neighboring base stations can cooperate to jointly transmit data or beamform to coordinate interference.Multi-cell cooperation is on the condition that the related BSs share CSI and user data, which brings huge signal overhead involving feedback overhead and backhaul overhead. When precoding algorithm is applied, tradeoff has to be made between the performance gain and the signal overhead. At the same time, how the shared information is available in practice has to be taken into account, for instance, the dedicated feedback channels which feedback the CSI from users to BSs, the backhaul channel with which the coordinated BSs exchange information, the channel duality and the channel estimation error, etc. So it is of great practical significance to explore distributed strategy to reduce the CSI sharing and the performance of coordinated precoding with limited feedback. With such a variety of forementioned factors, the dissertation makes an intensive study of multi-cell coordinated precoding in cellular system. The main content and contribution of this thesis can be summarized as:As the basis of CoMP technology, the thesis present in detail the classifying of CoMP from different point of view, the topology of the cellular system and the channel model of multicell coordination. It is classified by centralized coordination and distributed coordination according to if all base station are connected with a central processing unit, by full coordination and partial coordination depending on if all the BS cooperate with data-transmitting, by coordinated multi-cell transmission (base station coordinated transmission) and coordinated single cell transmission (coordinated beamforming/coordinated scheduling) relying on if the user data is shared between coordinated base stations. In simulation, the linear or hexagonal wyner model is often used to model the cellular system, and both small-scale fading and large-scale fading are to be considered. The channel model of coordinated multi-cell transmission is different from that of coordinated single cell transmission.(1) The precoding algorithms used in single cell are extended to multi-cell system and compared. Base station (BS) in different cells share in different extent both channel state information and data signals of their respective users to convert conventional interference to useful signals, which increases spectrum efficiency and greatly improves the quality of service of the users in cell edge region. In this paper, Block Diagnalization (BD) algorithm which designs proper precoding vectors in the null space of interfering channels in order to completely mitigate interference is applied in multicell joint transmission and compared with other algorithms, such as zero forcing algorithm, minimum mean square error algorithm, time division multiple access and single-user eigenbeamforming. The principles of interference mitigation and macrodiversity gain of BD are analyzed from the view of spatial dimension and the performance corresponding to each algorithm is discussed. Simulation results show that BD outperforms the algorithms aforementioned. Meanwhile, coordinated BD is superior to uncoordinated BD.(2) An optimal linear precoding scheme based on particle swarm optimization (PSO) is proposed for a multicell coorperation system. Particle swarm optimization (PSO) algorithm is simple in principle and easy to be realized. In addition, PSO, while with fast converge speed, has no specific requirements for the search space and the gradient information. In multi-cell coordinated transmission condition, we often aim to maximize the average spectrum efficiency under sum power constraint or per base power constraint. It finally converts to an optimization problem. Usually it is very difficult to validate the convexity of the objection function. In this dissertation, the PSO theory and its flow chart are presented and the convergence character is analyzed. Then the fitness value function of PSO adapting to multicell cooperation is derivated. With such a scheme, the optimal precoding vector could be easily searched for each user according to a simplified objective function. Simulation results show that the proposed scheme can obtain larger average spectrum efficiency and a better bit error rate (BER) performance than zero forcing (ZF) and minimum mean square error algorithm (MMSE).(3) Feedback and wireless backhaul overhead of base station coordination is one of the bottlenecks threatening its acceptance into commercial networks. To reduce the backhaul overheads and precoding complexity while benefiting from base station cooperation, the distributed precoding strategy based on maximizing the signal to leakage plus noise ratio (SLNR) is proposed. It only needs the local channel state information and the user data, which efficiently reduces backhaul overheads while enjoys the gain of the base station coordinated transmission. The distributed precoding strategy based on SLNR is compared with both fully coordinated zero forcing algorithm and fully distributed SLNR algorithm. Simulation results show that the proposed strategy get a good tradeoff between overheads and performance.(4) A precoding strategy of ICIN based on limited feedback is presented for a coordinated single cell transmission system. On the basis of inter-cell interference nulling (ICIN) at the transmitter sides of cooperative base stations, we propose an adaptive feedback bit allocation scheme to minimize the rate loss due to limited feedback. Furthermore, we investigate the scaling law of the total number of bits per user to maintain a constant rate loss. Simulation results show that the proposed feedback bits allocation strategy provides significant gain compared to equal bits allocation.(5) To reduce backhaul overhead and feedback overhead, a distributed precoding method based on signal-to-leakage-and-noise-ratio maximization for coordinated single cell transmission is proposed, which only needs the sharing of channel state information of interference channel and not the user data. It gets a better performance compared with eigenbeaming and intercell interference nulling which are two extreme sides. And then, to fully exploit feedback, we derive the expression of sum rate loss and find the optimal solution to minimize it by brute-force research. Simulation results show that the performance can be further improved through adaptively allocating bits relying on the relative strength between the desired and interfering signals within the cell.