Hybrid Scheduling in Heterogeneous Half Duplex and Full Duplex Wireless Networks

Abstract

Full-duplex (FD) wireless is an attractive 
communication paradigm with high potential for 
improving network capacity and reducing delay in 
wireless networks. Despite significant progress on 
the physical layer development, the challenges 
associated with developing medium access control 
(MAC) protocols for heterogeneous networks 
composed of both legacy half-duplex (HD) and 
emerging FD devices have not been fully addressed. 
Therefore, we focus on the design and performance 
evaluation 
of 
infrastructure 
scheduling 
algorithms 
for 
based heterogeneous networks 
(composed of HD and FD users). We develop the 
hybrid Greedy Maximal Scheduling (H-GMS) 
algorithm, which is tailored to the special 
characteristics of such heterogeneous networks and 
combines both centralized GMS and decentralized 
Q-CSMA mechanisms. Moreover, we prove that H- 
GMS is throughput-optimal. We then demonstrate by 
simple examples the benefits of adding FD nodes to 
a network. Finally, we evaluate the performance of 
H-GMS and its variants in terms of throughput, 
delay, and fairness between FD and HD users via 
extensive 
simulations. 
We 
show that in 
heterogeneous HDFD networks, H-GMS achieves 5
10× better delay performance and improves fairness 
between HD and FD users by up to 50% compared 
with the fully decentralized Q-CSMA algorithm. 
Index Terms—Full- duplex wireless, scheduling, 
distributed throughput maximization