Multi-hop Full-duplex Wireless Network Transmission Mechanism And Performance Evaluation Research

Full-duplex technology has been widely concerned by both academia and industry for its double throughput capacity and its potential to improve spectrum effectiveness.In this paper,we focus on the problem of multi-hop full-duplex wireless network transmission and performance evaluation mechanism.Firstly,we analyze several factors influencing full-duplex wireless network capacity gain.We construct the link scheduling mechanism of double-channel full-duplex wireless networks,and establish a comprehensive analytical framework to quantify the full-duplex network layer capacity gain.Results show that the network capacity of 1-D network is W/(nD)?4/(2+?),and the upper bound of 2-D network is 2/(2+?)?(1+?)r.Secondly,two kinds of link scheduling algorithms are proposed to optimize the network throughput in this paper.One is a centralized link scheduling algorithm(CLSA).In CLSA,we transform the link scheduling problem into finding the maximum independent set problem.We construct the interference graph of the multi-hop full-duplex wireless network,and then use the simulated annealing algorithm to find the maximum independent set.The other one is the distributed link scheduling algorithm(DLSA).In DLSA,we take asynchronous contention into consideration and modify the subtraction of backoff timer when it is non-zero to effectively utilize networks.Then,we implement these two algorithms,and make comparison with the ideal scheduling model and greedy link scheduling algorithm.Finally,in order to solve the issue of joint optimization of routing selection and scheduling,we use network flow to model and formulate it mathematically.Due to its NPhardness,we develop two routing algorithms.One is a minimal sum cost based routing selection algorithm(MinCostRo),and the other is a minimal maximum interference routing selection algorithm(MinMax Ro).After the route selection,we schedule the selected links.To evaluate the performance of our proposed algorithms,we make experiments and compare them with the shortest path routing algorithm(SPRA)and AODV-Based from different aspects on MATLAB.Results verify the effectiveness of MinCostRo and the throughput obtained by MinCostRo is 49% higher than SPRA and 33% higher than AODV-Based.