comparing arq and harq protocols for wsn and mimo · pdf filecomparing arq and harq protocols...

37
Digital Communications Semester Project On Comparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837 Supervisor: Marc Selig September, 2011

Upload: hoangxuyen

Post on 06-Feb-2018

227 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

Digital Communications Semester Project

On

Comparing ARQ and HARQ Protocols for WSN and

MIMO SYSTEMS

Student:

Praveen Francis Rego

Matriculation Number: 30104837

Supervisor:

Marc Selig

September, 2011

Page 2: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

2

Page 3: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

3

Declaration:

I hereby declare that this project work is my own work and has not been

submitted in any form for another degree or diploma at any university or

other institute of tertiary education. Information derived from the published

and unpublished work of others has been acknowledged in the text and a list

of references is given in the bibliography.

Praveen Francis Rego

Kassel, Germany

September, 2011

Signature…………

Page 4: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

4

Acknowledgement

I thank the Almighty God for providing me the knowledge and instilling in

me the confidence to perform all my undertakings successfully.

I am grateful to my supervisor Marc Selig, who has given me the opportunity

to work in this project and for all his help and guidance.

I am always grateful to my parents and relatives for their constant support

and encouragement at all times. Besides, I thank the entire comlab faculty for

their dedicated service in imparting high quality education and thereby

creating in me an unbound curiosity in specific fields of study. Furthermore, I

thank my dearest friends for their companionship, support, and

encouragement.

Page 5: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

5

Abstract:

A major concern in data communications is the effective and efficient control

of transmission errors caused by channel noise so that error-free data can be

delivered to the user. In view of this concern, to achieve error-free data,

various Data Link Layer error-control schemes were proposed. Automatic-

repeat-request (ARQ) schemes and Hybrid ARQ (HARQ) have been

considered.

The recent strides made in wireless communications and electronics have

encouraged the development of low-cost, low power, multifunctional sensor

nodes that are small in size and communicate reliably over short distances.

The tiny sensor nodes which consist of sensing, data processing, and

communicating components constitute the sensor network. Energy

conservation is one of the most important issues in wireless sensor networks

(WSN’s), where nodes are battery powered. The efficient transmission of data,

necessitate reliable and energy efficient error control schemes.

Although ARQ has been in place for many years and extensive studies exist

with respect to ARQ strategies primarily from the error correction point of

view, recent research investigate the integration of MIMO and ARQ so as to

exploit the best advantages of the combined schemes.

The aim of this project is to investigate and compare the performance of

existing ARQ and HARQ protocols applicable in wireless sensor networks

(WSN) and in multiple input multiple output (MIMO) systems.

Page 6: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

6

Contents Chapter 1 -Introduction .......................................................................................................... 7

1.1 Stop-and-wait (S-W) ARQ Scheme ................................................................................ 8

1.2 Go-back-N (GBN) ARQ scheme ..................................................................................... 9

1.3 Selective-repeat (SR) ARQ scheme ............................................................................... 9

1.4 Hybrid-ARQ (HARQ) scheme ....................................................................................... 10

Chapter 2- ARQ’S and HARQ’S in Wireless Sensor Networks ............................................... 13

2.1 Background and Motivation ....................................................................................... 13

2.2 Channel Aware Link Layer ARQ Protocol .................................................................... 13

2.3 Previous Related Work on ARQ protocols .................................................................. 14

2.4 Proposed channel aware ARQ protocol ...................................................................... 15

2.5 Cooperative and Non-Cooperative ARQ protocols for energy harvesting wireless

sensor nodes ..................................................................................................................... 18

2.6 Energy efficient adaptive error control (AEC-RSSI) Protocol ...................................... 21

Chapter 3- ARQ’s and HARQ’s Protocols in MIMO Systems ................................................. 25

3.1 Background and Motivation ....................................................................................... 25

3.2 MIMO single ARQ (MSARQ) and MIMO multiple ARQ (MMARQ) .............................. 25

3.3 HARQ-MIMO retransmission techniques ................................................................... 29

3.4 Cooperative Multicell ARQ in MIMO Cellular Systems ............................................... 32

Conclusion ......................................................................................................................... 35

Bibliography ...................................................................................................................... 36

Page 7: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

7

Chapter 1 -Introduction

The transmission of data over a channel introduces transmission errors. These

errors reduce the performance of the system. So, to achieve an error free

communication, error control has to be implemented to combat against the

transmission errors. Usually, error control is the task of the data link layer and

two error control schemes namely, automatic repeat request (ARQ) and forward-

error -correction (FEC) are mentioned [1].

The data before being transmitted over the channel is channel encoded.

Channel encoding is done to add redundant bits to the information. The

redundant bits provide robustness against the channel errors. This increases

the reliability of data transmission over the channel. In the ARQ

retransmission protocol, once the encoded data has been transmitted over the

channel the correctness of the code is checked by means of the parity check

bits appended to the data. If there is no error, the data transmitted is assumed

error free and a positive acknowledgment (ACK) is sent via the feedback

channel to the transmitter. The parity check bits are removed from the data

block and delivered to the user. However, in presence of an error, a

retransmission of the same data is requested again. This is notified to the

transmitter by means of a negative acknowledgement (NAK). Retransmission

continues via feedback channel until error-free data is received by the receiver.

Due to the simplicity and associated high reliability, ARQ schemes are widely

used in data communication systems [1].

In the forward-error-correction (FEC) scheme of error control, the encoded

data (codeword) has error-detection as well as error-correction capability. But,

when the error cannot be located and corrected, the erroneous codeword is

transmitted to the user. No retransmission is possible in this scheme and when

uncorrected errors are still present due to decoding failure, it is hard to

achieve high system reliability. ARQ schemes are often preferred over FEC

schemes for error control in data communication systems such as packet-

switched data networks and computer communication networks [1].

Page 8: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

8

In the subsequent sub-section a brief account of the various forms of the ARQ

schemes are mentioned. Depending on the way retransmission is carried out,

there are three basic types of ARQ schemes. Stop-and-wait (S-W), go-back-N

(GBN) ARQ and selective-repeat (SR) ARQ [1].

1.1 Stop-and-wait (S-W) ARQ Scheme

In [1], it is said that the stop-and-wait protocol is the simplest of the ARQ

schemes. The transmitter transmits the codeword, stops and waits (idling) for

an ACK/NAK from the receiver before it continues further transmissions.

This is shown in Fig. 1.1 below. In case a perfect transmission where there are

no errors, the receiver sends ACK via the feedback channel. On receiving the

ACK, the transmitter transmits the next codeword in the queue. But on

receiving NAK, the transmitter once again retransmits the same codeword.

Retransmissions continue until the transmitter receives an ACK. The

scheme although simple, has the drawback of being inefficient mainly because

of the time wasted in waiting for an ACK/NAK from the receiver. During this

time no data is transmitted and this results in low data throughput of the

system. Increasing the block length can be considered as a solution to reduce

the waiting time (idle time) for the ACK/NAK. But it also introduces more

errors due its increase in size. Moreover, due to constraints of data formats in

practical system, long code lengths are impractical to be considered.

Fig. 1.1. Stop-and-wait ARQ scheme [1]

Page 9: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

9

1.2 Go-back-N (GBN) ARQ scheme

In this scheme, the transmitter continuosly transmits the codewords without

waiting for the acknowledgement from the receiver. This is shown in Fig. 1.2

below. An ACK/NAK is received after one round-trip delay. Round-trip

delay is defined as the time elapsed between the transmission of a codeword

and the reception of ACK/NAK for that codeword. Thus, in one round-trip

delay a certain number of codewords N can be transmitted. When the receiver

encounters an error for a code word i, it deletes it and each of the succeeding

N-1 codewords. When the transmitter receives an NAK for a particular

codeword i, the control goes back to the codeword i and retransmits it until

the transmitter receives an ACK for it. This is then followed by the

retransmission of the succeeding N-1 codewords irrespective of whether they

were error-free or not. The major drawback of this protocol is that unnecessary

retransmissions of error free codewords following a codeword detected in

error results in reduced throughput and extra energy consumption [1].

Fig.1.2. Go-Back-N ARQ with N=7 [1]

1.3 Selective-repeat (SR) ARQ scheme

In the selective-repeat ARQ scheme, introduced in [1], the transmitter

continuosly transmits the codewords without waiting for the

acknowledgement from the receiver. This is shown in Fig. 1.3 below. The

receiver on detecting a transmission error, sends a NAK. On receiving the

Page 10: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

10

NAK for a particular codeword, the transmitter only retransmits the NAK’ed

codeword and continues with the transmission of other codewords in the

transmission buffer. The data received from all codewords at the receiver

must be ordered sequentially matching to the transmission sequence before

being sent to the user. A receiver buffer is needed to store the error-free

codewords in sequence following the codeword in error. On correctly

receiving the codeword in error on retransmission, the receiver buffer releases

any error-free codewords so as to sequence the transmitted codewords in

correct order. To avoid codewords being lost due to buffer overflow resulting

due to small buffer size, large receiver buffer storage is needed for the

selective repeat ARQ system [1].

Fig. 1.3. Selective-repeat ARQ [1]

1.4 Hybrid-ARQ (HARQ) scheme

It is seen from [1] that although the basic ARQ protocols provide error control

capabilities, they have several drawbacks. In order to overcome the

drawbacks, a hybrid scheme consisting of the combination of ARQ and FEC

were used and is referred to as hybrid ARQ. The FEC subsystem increases the

system throughput by correcting most of the frequently occurring error

patterns in the transmitted codewords. When a detectable but not correctable

error pattern is being realized, an ARQ retransmission is requested thus

Page 11: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

11

increasing the system reliability. As a result, the combination of FEC and ARQ

scheme provides higher reliability than a system operating only with FEC, and

higher throughputs than a system operating only with ARQ.

Hybrid ARQ schemes can be classified into two categories namely, type-1 and

type-2 schemes [1]. The type-1 schemes uses a code designed for error-

detection and error-correction. On receiving a codeword in error, the receiver

tries to correct the error depending on the error-correcting capability of the

code. If the number of errors is within the error-correcting capability of the

code then the errors are corrected and transmitted to the receiver or stored in

the receiver buffer. This constitutes the FEC action. If the error cannot be

corrected, the codeword is rejected and then the receiver asks for a

retransmission (ARQ action) of the same codeword until it is able to be

decoded correctly.

The type-2 schemes operate adaptively depending on the varying channel

conditions. In the first transmission, the message to be sent is encoded with

parity check bits having only error-detection capability. On detecting an error,

the receiver stores a copy of the error message in the receiver buffer and at the

same time asks for a retransmission. The codeword retransmitted is not the

original codeword but rather a block of parity check bits. The parity check bits

have error correcting capability chosen depending on the previous

erroneously transmitted codeword. When the parity check bits are received it

tries to correct the errors of the codeword stored in the receiver buffer. There

is no guarantee that the decoding will be successful and in the case of a

decoding failure, the receiver requests for second retransmission of the

NAK’ed codeword [1].

HARQ with soft combining is classified into (a) chase code combing and (b)

incremental redundancy combining.

(a) Chase code combining is a technique used to combine repeated data

packets at the receiver. The successive retransmitted data packets are identical

copies of original transmission having same code rate but have different

weights associated to each of them. Chase combining can be considered as

additional repetition coding and therefore has no coding gain. The weights are

Page 12: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

12

measure of the reliability of the packet. The decoder is a code combiner based

on maximum likelihood estimation of transmitted information. The soft

decisions statistics of previously received data packets are combined so as to

obtain the best estimate of the transmitted information packet and thus

improve the SNR of the received signal [11, 17].

(b) In the incremental redundancy scheme of soft decision combining, each of

the retransmitted data packets has different coded bits and different code rates

for the same set of information bits. They possess a coding gain since several

different codes are combined to form a lower rate code with stronger error

correction capabilities [11].

The type-1 scheme is different from soft chase combining with respect to the

storing/deleting of the erroneously received data packet. In chase combining,

the erroneous codewords when detected at the receiver are not deleted but

rather stored in the receiver buffer to combine with subsequent retransmissions

of the original data packet.

The type-2 scheme, though similar to the incremental redundancy soft code

combining technique, differs in the retransmission process. In type-2 scheme

the retransmitted codeword consists of only the parity check bits and not the

data. But in the incremental redundancy of code combining, the subsequent

retransmissions comprise of data along with different encoding bits each time.

Page 13: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

13

Chapter 2- ARQ’S and HARQ’S in Wireless Sensor Networks

2.1 Background and Motivation

The growth of wireless communications and the rapid progress in electronics

has encouraged the development of senor nodes that consume less power but

are able to guarantee efficient sensing and processing of data so as to bring

about reliable communications. With this as the objective, recent research is

based on designing efficient sensor networks [3].

Throughput and energy efficiency are important yardsticks to judge the

performance of a protocol [1, 4, 7, 9]. Most of sensor nodes are battery

powered and have low energy capabilities. Suitable protocols are designed so

as to achieve maximum throughput with high reliability of the transmitted

sensed information. The reliability of the transmitted link is prone to errors

due to random nature of the channel. So in order to ensure error-free

transmission of the link, data link error control schemes are used.

In the subsequent subsections the different error control protocols employed

in WSN’s shall be discussed highlighting their main features, operation and

bringing out the differences between them in terms of their achievable

performances.

2.2 Channel Aware Link Layer ARQ Protocol

In [4], the authors propose a channel aware link layer ARQ protocol. The

motivation behind the proposal is that the conventional ARQ protocols as

discussed in chapter 1 have no adaptive capabilities to track changes in

channel conditions. So, due to lack of channel state information, the

conventional ARQ methods have the tendency to waste energy on

unnecessary transmissions/retransmissions due to ill-conditioned channels.

Page 14: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

14

Thus the proposed channel-aware ARQ protocol aims to minimize the

unnecessary wastage of energy by inducing channel aware properties into

the ARQ transmission/retransmission process. In the paper, stop-and wait (S-

W) ARQ protocol has been used for the adaptive process. With the main

objective being to achieve energy efficient link layer communications, four

channel aware based protocols are mentioned. The protocols make use of the

channel state information incorporated in the feedback process to choose the

appropriate time instant for transmission/retransmission. In order to have

clarity of comparison between the conventional S-W, the related work already

done on channel probing, and the proposed protocols in [4], each of them will

be discussed briefly below.

2.3 Previous Related Work on ARQ protocols

The basic stop-and-wait (S-W) ARQ protocol or schemes possess no channel

checking capability, and transmission of a data frame is considered successful

only if the channel is good for the transmission period plus the ACK period. If

during the ACK period, the received signal strength is below a certain

predefined threshold power, the received data frame is assumed corrupted

and thereby rejected at the receiver.

The probing based S-W ARQ protocol is said to operate similar to the basic

S-W ARQ under good channel conditions. Probing implies the channel

tracking mechanism under bad conditions. When the channel is corrupted, on

receiving a corrupted data packet the protocol has the ability to send a NAK to

the transmitter. On receiving the NAK, the transmitter stops the transmission

of a new data packet and enters into the probing mode to check the quality of

the channel. Once an ACK for the channel condition is received, the

transmitter reverts back to transmitting mode to retransmit the data packet

again. Though not explicitly stated, it is assumed that the channel remains

good for the data transmission period following an ACK. The important

point to be considered is that the channel probing period (the duration for

which the probing is carried out) is not chosen optimally due to lack of

channel fading properties [4].

Page 15: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

15

2.4 Proposed channel aware ARQ protocol

The proposed fading channel aware ARQ scheme in [4] considers the channel

fading property and the current channel state to decide the correct time

instant for the transmission/retransmission of the data packet. The four

channel aware approaches are :

a) Channel-aware probing is based on average fading delay which is a

function of the receiver threshold power, AFD ( ): AFD is the average fading

duration for which the received signal remains below a certain threshold

power. The receiver on receiving a data frame notifies the transmitter whether

the received signal is above or below the preset receiver threshold power

level. This is accomplished via the ACK/NAK acknowledgement. On

receiving the NAK, the transmitter enters the probing mode until the received

signal power is greater than the receiver threshold power level. The receiver

threshold power is dependent on the signal fading margin and this affects

the fading dependent threshold power.

b) The second protocol ( is similar to but in addition to receiver

threshold power it makes use of the current signal level at the receiver and

transmits it to the transmitter via NAK. This protocol has the advantage of

being aware of the depth of fading channel and reporting it to the transmitter.

This additional information can be utilized in recovering faster from a deep

fade condition thereby improving the throughput and the energy efficiency.

c) The third proposed protocol ( ) utilizes the current signal level at

receiver in addition to its slope. The knowledge of slope characteristics

provides valuable information regarding the dynamics of the fading channel

to know whether it is improving or degrading further. The current channel

state and its past history obtained from the slope are sent via the NAK.

d) The fourth protocol is similar to the third protocol with the only difference

that the current signal level and its slope at the receiver are sent for both ACK

as well as NAK. So, on receiving an ACK, the transmitter does not

Page 16: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

16

immediately transmit a new data frame, rather it estimates for how long the

signal would remain good, and based on that decides on transmitting new

data or stays away from immediate transmission foreseeing bad channel

conditions. Also, based on the slope of the current signal level, a new data

frame transmission or probing will be initiated. The incorporated

modifications in the transmission timing decisions are expected to increase the

throughput, energy and delay efficiencies for the transmission process.

The simulation results in [4] clearly show the superiority in the performances

of the proposed fading channel-aware schemes over the basic S-W and the

probing schemes. The probing period for the probing scheme is fixed for two

time slots indicated by t = 2 in the figures below.

Fig.2.1. Throughput versus fading margin (transmit power) of [4]

The plot in Fig. 2.1 shows that has achieved significantly higher

throughput and the performance gain is more significant at moderate fading

margin. The plot of energy efficiency of compared in Fig. 2.2 indicate

that in general, a probing ARQ achieves a significant reduction in

unsuccessful attempts with respect to basic ARQ thereby increasing the

energy efficiency.

Page 17: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

17

Fig. 2.2. Energy efficiency performance of [4]

In [6], the authors propose another channel-aware adaptive ARQ. The paper

addresses the issue of inefficient utilization of the channel due to lack of

channel state information when the Berkley MAC (B-MAC) protocol is used.

B-MAC is a carrier sensing medium Access (CSMA) based protocol designed

for low power wireless sensor networks. B-MAC is a link layer protocol. It

assesses the channel by making use of the clear channel assessment (CCA)

functionality. CCA is used for tasks such as channel arbitration i.e. to check if

the channel is free or not, and for the acknowledgement of the link layer

frames to ensure reliability. The acknowledgement of link layer frames is

optional and the protocol can operate with or without ACK. The B-MAC

protocol has a bi-directional interface that allows the network services to

adaptively change the duty cycle of the B-MAC protocol depending on the

service needs. Although the B-MAC protocol is energy efficient with better

reliability than other MAC protocols, in scenarios where bursts of data have to

be transmitted reliably from different sensors to a destination (base station)

there occurs an inefficient utilization of the channel due to collision and loss of

ACK [6].

In [6], the authors propose an adaptive ARQ scheme for better channel

utilization and higher data rates. The adaptive error control is based on a fuzzy

logic control. The fuzzy logic control function selects the appropriate

Page 18: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

18

retransmission time instant for sending a particular data frame based on

signal strength of the nodes and distance between the sender and receiver

nodes. The performance of the proposed adaptive ARQ is compared with the

B-MAC protocol via simulation. Simulation results show the superiority of the

proposed adaptive ARQ protocol. The superiority in performance is expressed

in terms of higher delivery rate output.

Fig. 2.4. Data delivery rate (y) vs. Number of sensor motes (x) [6]

2.5 Cooperative and Non-Cooperative ARQ protocols for energy harvesting

wireless sensor nodes

The wireless sensor nodes in a sensor network need to be battery powered so

as to perform the tasks of sensing, processing, and communication. The

batteries may be either periodically replaced over time or the batteries may be

self–rechargeable i.e. recharge themselves by harvesting energy from the

surrounding environment. The energy harvesting technique provided is a

very attractive maintenance-free solution for low power sensor nodes [7].

In [7], the GAP4S system architecture as show in Fig. 2.5 was considered

where the location of sensor nodes are confined to a fixed distance

surrounding a power rich base station. The base station serves as an access

point and provides an extension to larger communication network. Each of the

Page 19: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

19

sensor nodes transmits data directly to the base station via the uplink wireless

channel. The energy harvesting or recharging is done via the downlink

microwave channel. The base station being rich in power has the

responsibility of scheduling the sensor nodes transmission time instants and

the relays to be chosen for each sensor node.

Fig. 2.5. GAP4S system architecture [7]

The main objective of the authors in [7] was to achieve maximum saturation

throughput from the sensor nodes to the base station while maintaining

reliable and fair transmission of the sensed information. Maximum saturation

throughput is defined in the paper as the maximum load that can be tolerated

by the sensor node without exceeding its energy harvesting rate. To fulfill the

second requirement of reliable transmission of the sensed information, ARQ

protocols were used. Two classes of ARQ protocols were considered, namely,

the conventional non-cooperative ARQ (ARQ-NC) protocol and the

cooperative ARQ protocol (ARQ-C).

The conventional ARQ protocol transmits the data directly to the base station

and waits for an acknowledgement. On receiving an ACK within the time

allotted for the ACK to be received, a new frame is transmitted. Otherwise the

same frame is retransmitted until the data frame is received correctly at the

base station.

Page 20: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

20

The Cooperative ARQ protocol (ARQ-C) makes use of the relaying

phenomenon where the sensor node chosen as the relay overhears the

information intended to the base station. It saves a copy of it in its buffer. On

receiving NAK for a particular date frame, it is the relay that has to retransmit

the data frame and not the source node. In doing so, unnecessary wastage of

energy due to retransmission from the source node is avoided. The relay offers

its energy in retransmitting the received NAK’ed data frame. To state it in

another way, the source sensor node borrows energy from the relay so that

its energy consumption rate does not exceed its recharging rate.

Fig. 2.6. Three node network scenario [7]

As mentioned earlier, it is the responsibility of the base station for choosing

the relay/relays for each sensor node for a data frame transmission. There are

many possibilities of choosing the relay for a sensor node. Either one sensor

node may act as relay for many source nodes or multiple sensor nodes can act

as relay for one source node. But in the latter case it is assumed that for every

data frame transmitted from a source node, a different relay would act each

time. It is said in [7] that using several relays for a single source node

improves the load and energy consumption balancing.

It was observed through simulation that for lower values of transmitted

energy per bit, the Cooperative ARQ protocol achieves a higher saturation

throughput than non-cooperative protocol. This is accomplished by the energy

borrowing scheme where energy deficient nodes ask neighboring energy rich

Page 21: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

21

nodes to retransmit the data frames. Thus the Cooperative ARQ protocol

achieves a saturation throughput which is twice as much as that of the non-

cooperative conventional ARQ when the energy is the limiting factor in the

system.

The authors in [7] have extended their work and proposed ARQ- protocol

in [8]. The protocol is a recursive version of the ARQ-C protocol where several

relays would be iteratively selected by the BS to overhear and store a copy of

the data frame intended to the receiver. The reason behind this proposal is

based on the assumption that the relays have a higher probability to

retransmit the data correctly compared to the previous relays or the source

node. Also, the relay being much closer to the receiver (BS) requires lesser

energy to retransmit the data. The simulation results in paper [8] show that

the saturation throughput of the ARQ- protocol is much better than both,

the ARQ-C and the non-cooperative ARQ protocols. However, the complexity

of the base station is slightly higher than the ARQ-C protocol.

2.6 Energy efficient adaptive error control (AEC-RSSI) Protocol

In [9], an energy efficient adaptive error control scheme is proposed. The AEC-

RSSI is a hybrid scheme comprising of the two data link layer error control

mechanisms namely ARQ and HARQ. Depending on the communication

distance between the sensor nodes, either of the two error control schemes is

adaptively chosen. When the transmitter sensor node receives NAK

information, it makes use of the information to calculate the location of the

receiver sensor node.

Energy efficiency is often considered as a metric for evaluating the

performance. To quantify and compare the performance of the AEC-RSSI

scheme with other schemes, first the energy efficiencies of the ARQ, FEC and

HARQ schemes are considered. Then energy efficiency performances are

compared by means of a mathematical approach. The comparison is done so

as to add clarity to distinguish the schemes mentioned later in the section.

Page 22: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

22

The transmission energy efficiency of WSN’s takes into consideration the

energy consumption and reliability of the transmission link. It is given by

(1- )….(2.1)

Where, is the energy consumption throughput, represents the link

layer frame error rate. The communication reliability can be expressed by

r = (1 - ) which represents the probability that a link layer frame is

successfully transmitted. r is further defined as ratio of energy consumed by

the link layer data frame ( to the total energy consumption ( . The

total energy comprises the energy needed to start/receive a data frame in

addition to the energy needed for coding/decoding depending on the error

control scheme. For the FEC scheme the decoding energy is included in the

total energy but for the ARQ scheme it is neglected.

The comparison of the energy efficiencies of the different schemes (FEC, ARQ,

and HARQ only) was considered for fixed length data frames and for variable

communication distances. The maximum number of retransmissions for ARQ

and HARQ was set to 1. In the figure below it is seen that the energy

efficiencies of all the schemes increased with the increase in payload size of

the data frame and for communication distances up to 40 meter. Within the 40

meter range, the ARQ scheme showed the best performance compared to FEC

and HARQ. This was because the additional decoding energy consumed by

the FEC and HARQ was far more than the energy consumed by ARQ due to

retransmissions. However, for distances greater than 40 meter, HARQ showed

better performance.

Page 23: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

23

a) Frame length=750 bytes. b) Frame length=1500 bytes.

. Fig. 2.7. Energy efficiency of FEC, ARQ, and HARQ [9]

The AEC-RSSI protocol in [9] comprises of the ARQ and HARQ error control

mechanism. The protocol makes use of the received signal strength indicator

(RSSI) for power control. Based on the RSSI the distance between the sensor

nodes is calculated. Depending on the distance, either ARQ or HARQ scheme

is adaptively chosen for error control depending on their respective

efficiencies. When the distance of communication is less than 40 meters, ARQ

is used due to its higher efficiency in this range. With distance greater and

equal to 40 meters HARQ is used for the error control of the lost link layer

frames.

In [9], it is shown the superiority in the performance of the proposed AEC-

RSSI scheme compared to the other schemes when the communication

distance is greater than 40 meters. Furthermore, the comparison in

performance is shown in Fig. 2.8 below. The benefits of the protocol was due

to the fact that AEC-RSSI protocol design has the inherent multiple

redundancy protection, which incorporates channel encoding technology and

diversity combining (soft combining) technique of retransmitting the link

layer frames.

Page 24: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

24

a) Length of link layer frame is 750 bytes . b) Length of link layer frame is 1500bytes.

Fig. 2.8. Comparison of AEC-RSSI, FEC, ARQ, and HARQ [9]

Page 25: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

25

Chapter 3- ARQ’s and HARQ’s Protocols in MIMO Systems

3.1 Background and Motivation

Research work carried out over the last decade has confirmed that by utilizing

multiple transmit and receive antennas multi-input multi-output (MIMO)

transceivers provide substantial spectral efficiency and diversity gain [2, 10,

16].

Although the diversity exploited through MIMO technique provides

substantial robustness to the errors introduced by the physical layer, practical

systems often encounter harsh, distortive or fading channels that can lead to

packet failure despite the robustness provided by MIMO diversity. So there

arises a necessity to provide additional protection to the transmitted signal

against channel errors. Error control schemes can be used to provide the

additional robustness thereby increasing the overall system throughput [16].

With the aim of improving the diversity provided by the ARQ scheme, several

antenna precoding schemes have been designed so as to increase the mutual

information throughput of the system. Single and multiple ARQ MIMO

protocols have been proposed and their performance is evaluated based on

linear and non-linear detection schemes [11, 13, 16].

In the subsequent sub-sections of this chapter, the various ARQ and HARQ

protocols of the data link layer in conjunction with the MIMO technique of the

physical layer will be explained.

3.2 MIMO single ARQ (MSARQ) and MIMO multiple ARQ (MMARQ)

The design of some MIMO systems is based on the BLAST (BELL Labs

Layered Space- Time) architecture, where substreams from the individual

Page 26: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

26

antennas are transmitted over different layers independently. BLAST is a

communication technique utilized in MIMO for achieving very high data rates

in rich scattering environments [2].

The encoding of the data can be done either by joint encoding or separate per-

antenna encoding. When the entire data packet is encoded and interleaved

before being demultiplexed into parallel data streams, the encoding is called

joint encoding. This is shown below in Fig. 3.1 (a).The joint encoding has the

limitation of increased complexity which grows with increase in number of

transmitter antennas. In per-antenna encoding the input data packet is

demultiplexed into parallel data streams and then each of the streams is

separately encoded and radiated from the corresponding transmit antenna as

shown in Fig. 3.1 (b) [11].

Fig. 3.1 (a) MIMO with joint coding. Fig.3.1 (b) MIMO with per-antenna coding [11].

In [11], a multi-user downlink packet data transmission is considered. The

base station is equipped with multiple antennas. In the joint encoding scheme,

the entire data packet has a single cyclic redundancy check (CRC) appended

to the data. The substreams being transmitted via different antennas

encounter different channel conditions and hence different error statistics.

When one or two substreams may be corrupted the entire data packet has to

be retransmitted owing to the single CRC for the entire packet. This leads to

Page 27: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

27

unnecessary retransmissions of the entire data packet reducing the

throughput efficiency of the system. Since the entire multiple transmit

antennas share a common single ARQ, the scheme of retransmission is called

MIMO single ARQ (MSARQ).This is shown in Fig. 3.2 (a) below.

To overcome the drawback of the MSARQ scheme, per-antenna encoding with

multiple ARQ processes has been proposed. A single CRC is attached to each

substream and detection of each substream is done independently. Thus only

those substreams which are corrupted will be retransmitted from that

particular transmitter antenna. The scheme of retransmission is called MIMO

multiple ARQ (MMARQ) and is shown in Fig. 3.2 (b) below.

Fig. 3.2. Transmitter structures for MSARQ AND MMARQ scheme [11]

In [11], the receiver side functionality consists of the HARQ combining and

the interference cancellation process. MMARQ and the MSARQ process

operate differently. In MMARQ, the HARQ combining technique and

interference cancellation step are blended advantageously. The content of the

detected substreams are validated by means of their cyclic redundancy check.

If the cyclic redundancy check detects uncorrectable error, a copy of the

corrupted stream will be stored and a retransmission will be asked for that

particular stream. A soft decision combining of all the received packets is done

so as to increase the probability of correct symbols being received. This is the

HARQ combining technique. The data obtained from HARQ combining is

Page 28: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

28

then used for the interference cancellation. This ensures that the interference

contribution of the data stream is removed.

However, in MSARQ scheme the combination of packet combining and

interference cancellation cannot be done easily. Firstly each data stream is

decoded and interference cancellation is done. The data streams are then

demulitplexed into the single transmitted stream and correctness of the data

stream is then validated by means of the single CRC .The single CRC check is

ineffective in the sense that there may be several data streams which are

erroneous but have already undergone interference cancellation prior to CRC.

This result in a decoding error of each substream to propagate to substreams

decoded later. Based on the CRC, a retransmission of corrupted data frame is

requested and HARQ soft combining is performed.

MMARQ and MSARQ substream error performance carried out through

simulation in [11] show the superiority of the MMARQ scheme. The

performance of the MMARQ scheme achieves 26-40% throughput

improvement compared to MSARQ in case of perfect channel state and perfect

uplink feedback. However, comparatively a better performance of the

MMARQ scheme is achieved in case of non-ideal environment and non

perfect channel state information. MMARQ achieves a throughput

improvement of 30-45%. It is seen that MMARQ is relatively less sensitive to

estimation noise and feedback error.

In [13], a spatially multiplexed MIMO system is considered with channel state

information only at the receiver .Two types of HARQ schemes are considered,

namely single HARQ and multiple HARQ. In single HARQ the transmission

and retransmissions from all the layers is controlled by a single HARQ while

in multiple HARQ, the packets are preassigned to each antenna and

transmission/retransmissions take place only from the preassigned antenna.

Time division multiplexing is used for data transmission. To prevent

unnecessary retransmissions of entire data packet in single HARQ each time

slot is divided into a number of independent subtime slots.

Page 29: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

29

Joint and separate detection was used for each of the HARQ schemes. In joint

detection the symbol received in current time slot and the retransmitted

symbols in subsequent time slots are jointly detected.

In [12], a joint detection combing scheme for single HARQ was proposed. The

joint detection was based on combing the soft decisions at receiver output

either prior to interference cancellation (pre-combining) or after the

interference cancellation (post-combining). When the detection was carried

out for linear zero-forcing (LZF) and the minimum mean-square error

(LMMSE) receivers, the precombining based HARQ achieved better

throughput gain than the post-combing HARQ scheme.

Similarly in [13], by virtue of simulation, it was seen that for a certain SNR in

joint and separate detection process, single HARQ achieves higher throughput

gain than multiple HARQ for a linear receiver structure. The superiority in

single HARQ is due to the random interleaving across the layers prior to

transmissions, thereby exploiting the diversity to develop robustness against

channel errors and this facilitates the detection process. But when the non-

linear interference cancellation (detection) is used by virtue of V-BLAST

architecture, multiple HARQ always outperforms the single-HARQ. Here, the

gain due to interference cancellation is far more than the gain achieved due to

interleaving across layers in case of single HARQ.

3.3 HARQ-MIMO retransmission techniques

In [15], the authors propose a cross layer design MIMO-HARQ protocol which

jointly exploits the spatial diversity of the MIMO transmission at the physical

layer and the link reliability of the data link layer. The proposed protocol

makes use of the MIMO V-BLAST architecture so as to benefit from the error

control performance in the link layer. The HARQ error control scheme is

adopted with selected repeat (SR) ARQ as the retransmission protocol.

Page 30: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

30

The transmission of data packets is based on multiplexing the encoded

codeword of certain block length into sub-blocks via each of the several

transmitting antennas. This is shown in Fig. 3.3 below. Depending on the

retransmission technique adopted the cyclic redundancy check may or may

not be appended to each sub-block to validate the content. If the error pattern

can be corrected by the error-correcting code, then no retransmission is

required. Thus the transmission scheme by exploiting the spatial diversity

develops robustness against transmission error and at the same time

minimizes the number of retransmissions, in turn making better utilization of

resources.

Fig. 3.3. Dividing HARQ codewords over the transmit antennas [15]

Failure Block transmission rate (FBR) is a parameter which determines the rate

of failed transmission and it is considered as an important quality of service

criteria. By prioritizing the retransmission process the block error rate (BLER)

can be minimized. In [15], to ensure high reliability to the retransmitted data

various techniques are considered and are briefly discussed below.

In the basic retransmission technique, no CRC is used for each sub-block. Even

due to single error the whole block must be retransmitted. In any case the

entire block must be retransmitted, so the need for prioritizing the

retransmissions to new data transmission is no longer valid.

Page 31: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

31

In the Alamouti retransmission (AR) technique, the CRC is not appended to

each sub-block and on detecting an error, the entire data block is

retransmitted. Priority is give to the retransmission process where each

symbol of the retransmitted block is coded by the Alamouti coding scheme

([15], referenced to [10] within the paper).

Another retransmission technique is based on resending the entire block on

that antenna having the best channel conditions. This is called the transmit

antenna selection retransmission technique (TASR). It is the responsibility of the

receiver which has perfect channel state information (CSI) to feedback the CSI

to the transmitting antenna.

The third technique (AR+CRC) facilitates either an entire block or a sub-block

retransmission. The sub-block retransmission is made possible by the presence

of the CRC appended to the end of each sub-block data. Due to the individual

CRC’s there is added complexity to the buffering and processing.

The fourth technique (TASR+CRC) is also a combination of the best antenna

selection with CRC appended to each sub-blocks. Retransmission of a single

sub-block as well as entire block is possible.

Simulation results carried out in [15] validate the fact that the proposed

HARQ-MIMO retransmission techniques achieve better throughput and

minimize the block error rates. The Alamouti coding scheme achieves the

better performance compared to the transmit antenna selection scheme but at

the expense of added complexity. Also individual CRC’s appended to each

sub-block results in added complexity but in very little additional

performance gain.

Page 32: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

32

3.4 Cooperative Multicell ARQ in MIMO Cellular Systems

In [14], multiple base stations are considered and each of the base stations is

equipped with multiple antennas. The cooperation among the base stations

helps in achieving better coverage and throughput in the MIMO cellular

system. The enhanced performance due to cooperative multipoint processing

(CoMP) is virtue of exploiting the spatial diversity obtained by means of the

broadcast nature of the wireless channel. With the objective of increasing the

reliability of the transmission links, by exploiting the broadcast nature of the

wireless transmission, ARQ protocol at the data link layer is combined with

the cooperative diversity at the physical layer.

In [14], a three-cellular uplink transmission is considered with one serving

base station and two relaying base stations. This is shown in Fig. 3.4 below.

The serving base stations are connected to its neighbors via wired back haul

links. The back haul links serve the two purposes. Firstly, the cooperation

among the base stations to facilitate better transmission performance is done

via back haul links. Secondly, it serves as pathway to increase the reliability of

the transmission link by facilitating the retransmission of the data packets

when the data packet received at the serving base station is corrupted.

Fig.3.4. CoMP for 3 cells [14]

Page 33: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

33

To enhance the reliability of the transmission links, conventional ARQ

protocols and cooperative multicell ARQ protocols are considered. The

conventional ARQ protocol does not make use of the relaying neighboring

base stations and fails to exploit the diversity achievable through relaying.

The cooperative multicellular ARQ protocols are of three types namely,

opportunistic decode-and-forward (ODF), amplify and forward (AF) and compress

and forward (CF).

In ODF the neighboring BS’s overhear the information intended to the serving

BS and try decoding it. On successfully decoding it re-encodes the data again

and forwards it to the serving BS. Maybe, due to the failure of the link

between the relays and serving BS (destination), a NAK is sent to the source

terminal as well as to the BS’s involved in relaying. The communication for

retransmission is carried out via the back haul links between the cooperating

BS’s and the serving BS’s.

In AF protocol, all the neighboring multiantenna base stations take part in

retransmitting their information from each of their antennas. The AF protocol

does not decode the information rather only amplifies and forwards the

information. The absence of decoding makes the AF scheme less complex in

comparison to the ODF scheme of retransmission.

In the CF scheme the information received at the base station is compressed

and forwarded to the base station. The performance of the scheme improves

with the increase in the reliability of the link between the relay and the serving

base station.

In [14], the comparison between the conventional ARQ and the CoMP ARQ

protocols were based on the average packet error rate (PER) for each of the

protocols. The packet error rate depends on the instantaneous SNRs of the BS

being served and also on the effective instantaneous received SNR at serving

Page 34: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

34

BS via the neighboring BS. Simulation results carried out show that all the

cooperative multicellular ARQ protocols provide much lower average PER.

Also the performance gap between each cooperative mutlicell ARQ protocol

reduces with increase in the SNR of the back haul links. Additionally, the AF

protocol performs the best compared to the other two protocols when the

reliability of the backhaul link is considered to be at a certain minimum. Also,

the performance of the proposed protocols improves with increase in number

of BS’s and with number of antennas at each BS.

Page 35: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

35

Conclusion

In this project a literature survey has been conducted on the application of

error correction schemes in wireless sensor networks and in MIMO systems.

Throughout the survey, the significant benefits of exploiting the cross layer

analysis of the physical layer and the data link layer has been observed. The

data link layer error control schemes in conjunction with the physical layer

architecture of wireless sensor networks and MIMO have considerably

improved the overall system performance that can be achieved. The gain in

system performances are quantified based on certain common performance

yardsticks such as throughput and energy efficiencies. In particular, special

emphasis was laid on surveying the various ARQ’s and HARQ’s in wireless

sensor networks and in MIMO systems. Different protocols were compared

and contrasted, highlighting their performance characteristics. The remarkable

benefits of the cross layer analysis have fostered a continued research interest

in this specific field of study.

Page 36: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

36

Bibliography

[1] Shu Lin, Costello D, Miller M, "Automatic-repeat-request error-control

schemes," Communications Magazine, IEEE , vol.22, no.12, pp.5-17, December 1984

[2] Wolniansky, P.W.; Foschini, G.J. ; Golden G.D.; Valenzuela, R.A.; "V-BLAST: an

architecture for realizing very high data rates over the rich-scattering wireless

channel," Signals, Systems, and Electronics, 1998. ISSSE 98. 1998 URSI International Symposium

on, vol., no., pp.295-300, 29 Sep-2 Oct 1998

[3] Akyildiz, I.F.; Weilian Su; Sankarasubramaniam, Y.; Cayirci, E.; "A survey on sensor

networks," Communications Magazine, IEEE, vol.40, no.8, pp. 102- 114, Aug 2002

[4] Swades De; Cavdar, H.D.; "Channel-Aware Link Layer ARQ Strategies in Wireless

Networks," Wireless Communications and Mobile Computing Conference, 2008. IWCMC '08.

International, vol., no., pp.389-394, 6-8 Aug. 2008

[5] Ke Cheng; Srinivasan, S.M.; Tripathi, A.; "Adaptive ARQ in wireless sensor

networks," Computer Science and Information Technology (ICCSIT), 2010 3rd IEEE International

Conference on, vol.5, no., pp.181-184, 9-11 July 2010

[6] J. Polastre, J.Hill and D. Culler, (2004) “Versatile Low Power Media Access for Wireless Sensor

Networks”, Sensys’04, November 3-5

[7] Marco Tacca, Member, IEEE, Paolo Monti, Member, IEEE, Andrea Fumagalli, Member,

IEEE,(2007) “Cooperative and Reliable ARQ Protocols for Energy Harvesting Wireless Sensor Nodes”

[8] Tacca, M.; Monti, P.; Fumagalli, A.; “Cooperative and Reliable ARQ Protocols for Energy

Harvesting Wireless Sensor Nodes," Wireless Communications, IEEE Transactions on, vol.6, no.7,

pp.2519-2529, July 2007

[9] Yong Jin; Jinyi Chang; Deguang Le; “A High Energy Efficiency Link Layer Adaptive Error

Control Mechanism for Wireless Sensor Networks," Computational Intelligence and Software

Engineering (CiSE), 2010 International Conference on, vol., no., pp.1-4, 10-12 Dec. 2010

[10] G. J. Foschini, “Layered space–time architecture for wireless communication in a fading

environment when using multiple antennas,” Bell Labs Tech. J., vol. 1, no. 2, pp. 41–59, 1996

[11]Haitao Zheng, Angel Lozano, Mohamed Haleem, “Multiple ARQ processes for MIMO Systems”

EURASIP Journal on Applied Signal Processing ; 772–782, 2004

[12] Onggosanusi, E.N.; Dabak, A.G.; Yan Hui; Gibong Jeong; “Hybrid ARQ transmission and

combining for MIMO systems," Communications, 2003. ICC '03. IEEE International Conference

on, vol.5, no., pp. 3205- 3209 vol.5, 11-15 May 2003

Page 37: Comparing ARQ and HARQ Protocols for WSN and MIMO · PDF fileComparing ARQ and HARQ Protocols for WSN and MIMO SYSTEMS Student: Praveen Francis Rego Matriculation Number: 30104837

37

[13] Bai, C.; Krzymien , W.A.; Fair, I.J.; “Hybrid-arq for layered space time MIMO systems with

channel state information only at the receiver," Communications, IET, vol.4, no.14, pp.1765-

1773, September 24, 2010

[14] Woradit, K.; Quek, T.Q.S.; Lei, Z.Z.; "Cooperative multicell ARQ in MIMO cellular

systems," Signal Processing Advances in Wireless Communications (SPAWC), 2010 IEEE Eleventh

International Workshop on , vol., no., pp.1-5, 20-23 June 2010

[15] Ajib, W.; Haccoun, D.; Frigon, J.-F.; , "Data link control for multiple input multiple output

wireless systems," Wireless Telecommunications Symposium, 2005 , vol., no., pp. 112- 119, April

28-30, 2005

[16] Zhihua Shi; Haitong Sun; Chunming Zhao; Zhi Ding; , "Linear precoder optimization for ARQ packet retransmissions in centralized multiuser MIMO uplinks," Wireless Communications, IEEE Transactions on , vol.7, no.2, pp.736-745, February 2008 [17] Chase, D.; , "Code Combining--A Maximum-Likelihood Decoding Approach for Combining an Arbitrary Number of Noisy Packets," Communications, IEEE Transactions on , vol.33, no.5, pp. 385- 393, May 1985