Transceiver Design For Two-Way Or Multi-Way Relay Networks

Cooperative relaying is widely investigated due to its potential of enhancing the coverage and improving the capacity and reliability of wireless communication systems. However, the conventional one-way relaying suffers from significant spectral loss under the half-duplex constraint. Combining with physical layer network coding, two-way re-laying overcomes this drawback and has attracted considerable attention. Besides, multi-way relaying is drawing more and more attentions with the growing of the emerging applications, such as video conference, online games as well as disaster and emergence relief. This dissertation focuses on the typical two-way or multi-way relaying scenar-ios, i.e., multi-pair two-way relay systems, multi-cell multi-user two-way relay systems, multi-antenna multi-way relay systems and distributed multi-way relay systems, and in-vestigates the transceiver designs in each scenario.In this dissertation, a relaying protocol is first developed based on analogue network coding to avoid inter-pair interference in multi-pair two-way relay networks. Power al-location for users and the relay is optimized to maximize the achievable sum rate of the network. Besides, a simple and effective user pair selection method is integrated in the design to enhance the system sum rate.Then a general multi-cell multi-user two-way relay network is considered in the dis-sertation. This network model includes the thoroughly studied two-way relay networks, multi-pair two-way relay networks and multi-user two-way relay cellular networks, as special cases. A transceiver design is proposed for such a network under interference free constraints. The power allocation at the relay, base stations (BS) and users is investigated to maximize the system sum rate. The sum rate performance of the relaying scheme is finally evaluated through numerical examples.The above work focuses on improving the system sum rate and does not consider the traffic of each user. Basically, a user with much traffic should be allocated with more power from the point of satisfying the transmission demand of each user, and a user with channels of good quality should be allocated with more power from the point of improving the system sum rate. The dissertation studies the power allocation schemes for both regenerative and non-regenerative multi-way relay networks. Considering dif-ferent traffic demands of each user, a demand metric is defined as the product of a user's throughput and its queue length. Then a queue-aware power allocation scheme is pro- posed to stabilize the queue of each user by maximizing such a sum demand metric of all the users. With the scheme, the system throughput performance is improved while keeping all queues relatively stable.The aforementioned schemes are all realized via a multi-antenna relay. However, in some networks such as wireless sensor networks, the nodes are usually equipped with single antenna and multiple single-antenna relays form a a virtue antenna array. This work studies the schemes of such distributed relaying networks. Two distributed relaying schemes are proposed based on two widely adopted criteria, i.e., zero-forcing relaying and minimum mean square error relaying. For the distributed zero-forcing relaying, the precoding vector at the relays lies in the nullspace of the inter-pair interference. Through the optimization of the precoding vector, the ratio of the total useful signal power to noise power is maximized. For the distributed minimum mean square error relaying, the precoder at the relays and the receive filters at the users are jointly optimized.