05-09-2012, 03:54 PM
Channel Access Delay and Buffer Distribution of Two-User Opportunistic Scheduling Schemes in Wireless Networks
[attachment=33454]
Abstract
In our earlier works, we proposed rate adaptive
hierarchical modulation-assisted two-best user opportunistic
scheduling (TBS) and hybrid two-user scheduling (HTS) schemes.
The proposed schemes are innovative in the sense that they
include a second user in the transmission opportunistically using
hierarchical modulations. As such the frequency of information
access of the users increases without any degradation of the
system spectral efficiency (SSE) compared to the classical opportunistic
scheduling scheme. In this paper, we analyze channel
access delay of an incoming packet at the base station (BS) buffer
when our proposed TBS and HTS schemes are employed at
the BS. Specifically, using a queuing analytic model we derive
channel access delay as well as buffer distribution of the packets
that wait at BS buffer for down-link (DL) transmission. We
compare performance of the TBS and HTS schemes with that of
the classical single user opportunistic schemes namely, absolute
carrier-to-noise ratio (CNR)-based single user scheduling (ASS)
and normalized CNR-based single user scheduling (NSS).
INTRODUCTION
ALTHOUGH the classical single-best user opportunistic
scheduling (see, for example, [1], [2]) offers a considerable
overall throughput gain (which is known as multiuser
diversity gain), it suffers from low frequency of information
access. Specifically, with higher number of users in the
system, information access probability of a given user may
decrease. This results in higher channel access delay of the
packets that are waiting at the base station (BS) buffer to be
transmitted. The channel access delay of a packet waiting at
the BS buffer is important since with a lower access delay, a
receiver starts to receive and decode information bits earlier.
This has motivated us to design two-user, namely, two-best
user scheduling (TBS) and hybrid two-user scheduling (HTS)
schemes, which are presented in [3], [4], respectively. Brief
description of these schemes will be given later in this paper.
These were modulation-assisted solutions for the two-user
opportunistic scheduling.
NUMERICAL RESULTS
For our numerical results we consider both i.i.d. and i.n.d.
fading environments. We consider a system of 5 users and
maximum constellation size of 1024 which corresponds to
maximum transmission rate of 10 packets per slot. We assume
that users have a target BER of 10−4. We use MATLAB
optimization toolbox to find the rate switching thresholds and
statistical toolbox to generate independent Rayleigh fading
channel gains for the users. We assume Bernoulli packet
arrival process to the users’ buffers at the BS.
CONCLUSION
In this paper, we have analyzed the buffer distribution and
CDF of delay of our earlier proposed rate adaptive modulationassisted
TBS and HTS schemes. We have also compared these
performances with those of the classical single user scheduling
schemes. Some selected numerical results showed that for
an i.i.d. fading environment for Bernoulli arrival the TBS
outperforms the classical SBS in terms of buffer distribution
and delay distribution. In other words, the probability that an
incoming packet will get access to the channel within
[attachment=33454]
Abstract
In our earlier works, we proposed rate adaptive
hierarchical modulation-assisted two-best user opportunistic
scheduling (TBS) and hybrid two-user scheduling (HTS) schemes.
The proposed schemes are innovative in the sense that they
include a second user in the transmission opportunistically using
hierarchical modulations. As such the frequency of information
access of the users increases without any degradation of the
system spectral efficiency (SSE) compared to the classical opportunistic
scheduling scheme. In this paper, we analyze channel
access delay of an incoming packet at the base station (BS) buffer
when our proposed TBS and HTS schemes are employed at
the BS. Specifically, using a queuing analytic model we derive
channel access delay as well as buffer distribution of the packets
that wait at BS buffer for down-link (DL) transmission. We
compare performance of the TBS and HTS schemes with that of
the classical single user opportunistic schemes namely, absolute
carrier-to-noise ratio (CNR)-based single user scheduling (ASS)
and normalized CNR-based single user scheduling (NSS).
INTRODUCTION
ALTHOUGH the classical single-best user opportunistic
scheduling (see, for example, [1], [2]) offers a considerable
overall throughput gain (which is known as multiuser
diversity gain), it suffers from low frequency of information
access. Specifically, with higher number of users in the
system, information access probability of a given user may
decrease. This results in higher channel access delay of the
packets that are waiting at the base station (BS) buffer to be
transmitted. The channel access delay of a packet waiting at
the BS buffer is important since with a lower access delay, a
receiver starts to receive and decode information bits earlier.
This has motivated us to design two-user, namely, two-best
user scheduling (TBS) and hybrid two-user scheduling (HTS)
schemes, which are presented in [3], [4], respectively. Brief
description of these schemes will be given later in this paper.
These were modulation-assisted solutions for the two-user
opportunistic scheduling.
NUMERICAL RESULTS
For our numerical results we consider both i.i.d. and i.n.d.
fading environments. We consider a system of 5 users and
maximum constellation size of 1024 which corresponds to
maximum transmission rate of 10 packets per slot. We assume
that users have a target BER of 10−4. We use MATLAB
optimization toolbox to find the rate switching thresholds and
statistical toolbox to generate independent Rayleigh fading
channel gains for the users. We assume Bernoulli packet
arrival process to the users’ buffers at the BS.
CONCLUSION
In this paper, we have analyzed the buffer distribution and
CDF of delay of our earlier proposed rate adaptive modulationassisted
TBS and HTS schemes. We have also compared these
performances with those of the classical single user scheduling
schemes. Some selected numerical results showed that for
an i.i.d. fading environment for Bernoulli arrival the TBS
outperforms the classical SBS in terms of buffer distribution
and delay distribution. In other words, the probability that an
incoming packet will get access to the channel within