A Performance Comparison Between CACs and
SCACs Based Topology-Transparent Scheduling
Yijin Zhang, Jun Wei, Ming Zhang, Aijie Zou and Feng Shu
School of Electronic and Optical Engineering
Nanjing University of Science and Technology
Nanjing, China
Abstract—Topology-transparent sched uling is a deterministic
medium access control protocol for mobile ad hoc networks and
wireless sensor networks. This approach only depends on two pa-
rameters: the number of nodes in the network and the maximum
size of neighborhood. An attractive property is that each node
can guarantee a bounded delay in wh ich several packets can be
successfully transmitted, no matter which nodes are its neighbors.
In this paper, we study conflict-avoiding codes (CACs) and
strongly conflict-avoiding codes (SCACs) based scheduling for
slot-synchronized and asynchronous network, respectively, since
it i s difficult for mobile nodes to achieve frame-synchronization.
The effect of synchronization on maximal delay is examined.
In addition, we investigate average delay and average energy
consumption by simulation.
Index Terms—Top ology-transparent scheduling, conflict-
avoiding codes, strongly conflict-avoiding codes, synchronization.
I. INTRODUCTION
In mobile ad hoc networks (MANETs) or wireless sensor
networks (WSNs), it is challenging for medium access control
(MAC) protocols to provide bounde d delay and short- te rm
fairness, due to the nod e mobility and multi-hop communi-
cation. As such, topology-transpare nt strategies [1]–[6] have
been drawing attentions to overcome these problems. This
approa ch only depends on two param eters: the number of
nodes in the network, and the maximum size of neighborhood.
A topology -transparent schedule (TTS) is said to be g-robust
if it ensures a favorable property that each node can send g
collision-fre e packets to its each neig hbor within a predefined
schedule length . This is a strict guarantee with pro bability one,
in con trast to random and contention-based schemes.
A TTS is acco mplished by assigning each node a determin-
istic binary sequence. Ea ch node transmits a packet within
one time slot duration if the sequence value is one, an d
keeps listening for one time slot if it is zero. Th e transceiver
is half-duplex, i.e., a tran sm itting node cannot receive any
message. A collision occurs if a packet starts or en ds within the
transmission duration of another packet. Only those packets
without suffering any co llision c an be rece ived correctly.
Considering different synchron iz ation level, the re are d-
ifferent combinatorial objects underling the corresponding
TTS constructions. By assuming that the nodes a re frame
synchro nized, most known TTSs [1], [2], [5], [6] are de-
signed based on cover-free families and orthogonal arrays.
In [3], [4], optica l orthogonal codes (OOCs) are employed
to design TTSs in the slot-synchronized and asynchronous
network. However, the Hamming auto-c orrelation constraint in
OOCs is inessential in TTSs to avoid con flicts. This motivates
researchers to find that conflict-a voiding codes (CACs) [8],
[9] and strongly con flict-avoiding codes (SCACs) [10] are
more su itable to construct 1-robust TTSs for these two weak-
er synchronization models, as their combinatorial structures
only depe nd on Hamming cross-co rrelation. Moreover, they
both have the minimum Hamming weight, which yields re-
transmitting times and energy cost as small a s possible.
Three criteria are commonly conside red to eva luate the
perfor mance of TTSs: (i) the maximal delay, w hich is upper
bounded by the schedule length; (ii) the average delay, which
depends on the sequence structure; (iii) the energy consump-
tion. In general, the energy consumed in listening state is equal
to 50%-80 % [7] of that in transmitting state.
In this pap er, w e will generalize CACs and SCACs to
support g-robust property for any g, and further examine the
effects of synchronization on the maximal delay, average delay
and average energy consumption. A similar study on OOCs
based TTSs can be found in [4], but only focused on the
maximal delay.
The rest of this paper is organized as follows. After setting
up some notation and definitio ns in Section II, we in Section
III present a sufficient combinatorial condition of g-robust
TTSs for each synchronization level. Section IV compares
lower bounds on the minimum schedule leng th, and Section V
compares maximal delays by CAC and SCAC constructions.
Section VI provides a performance study by simulation in
terms of average delay a nd average energy consumption. Fi-
nally, we close in Section VII with some concluding remarks.
II. NOTATION AND DEFINITIONS
For i = 1, 2, . . . , K, the binary sequence associated with
user i is specified by a row vector
s
i
:= [s
i
(0) s
i
(1) . . . s
i
(L − 1)],
and its Hamming weight is defined as w(s) :=
P
L−1
t=0
s(t).
A sequence can also be defined as the set of all time indices
in a period where the value of the protocol sequence is equal
to 1. Let Z
L
= {0, 1, . . . , L−1} be the set of integers reduced
modulo L. A subset I of Z
L
is associated with a binary
sequence s(t) of length L, by setting s(t) = 1 if and only
if t ∈ I.
978-1-4673-7687-7/15/$31.00 ©2015 IEEE