Research On The Performance Of Wireless Multi-hop Network

With the popularity of mobile Internet, wireless network services convert to provide high-speed data access from the current voice services, and the traditional single-hop network is facing a series of challenges. While mesh, relay and other distributed wireless communication technology has developed, the next generation mobile communication system will be a distributed wireless network with support of the infrastructure network, and combines the advantages of the cellular mobile communication and the distributed wireless communication, providing a management, operations, low-cost mobile broadband communication system for telecom operators.In this thesis, focused on studying the wireless multi-hop networks based on cellular networks, and the networks are divided into two types of mesh mode and relay mode. So the contents of study include two parts.In the first part, the model of mesh mode wireless multi-hop network is created, and a system level simulation platform is completed by C language. Based on this basis, link scheduling algorithms LS-MTNT and LS-MMET are designed. The total network throughput will be maximized when the links scheduled by LS-MTNT, and the minimum throughput will be maximized when the links are scheduled by LS-MMET. Finally, the algorithms are accessed by simulations. The minimum end-to-end throughput can achieve the 80% of its upper bound when links are scheduled by algorithm LS-MTNT. Compared with greed algorithm, the gain of node end-to-end throughput is more than 50% when the links are scheduled by LS-MMET.In the second part, the model of LTE-A Relay System which is typical representative of relay mode wireless multi-hop network is created, and construct the interference model and resource allocation model of both in-band and out band for UE and relay. A system level simulation platform is competed by C language. Finally, the network performance of Case 1 and Case 3 defined in 3GPP are measured by SINR and throughput. The performance differences of different scenarios are quantitatively analyzed, and the gains are also analyzed compared the scenario of eNode only.