PACKET SPEECH INTERPOLATION IN MOBILE TELEPHONE SYSTEMS A thesis by lindayy

VIEWS: 1 PAGES: 6

More Info
									Pi
Z<A 5M
PACKET SPEECH INTERPOLATION IN
MOBILE TELEPHONE SYSTEMS
A thesis submitted in fulfilment of the requirement for the degree of
DOCTOR OF PHILOSOPHY
in the
Department of Electrical Engineering, The University of Adelaide
by
B.Sc., B.E.(Hons)
JOHN CHARLES ELLERSHAW,
June 1979
TABLE OF CONTENTS
(v)
Abstract
(vii)
(viii)
Statement
Acknowledgements
(ix)
List of Symbols
Introduction
1.
1.1
Background
1
1.2 Packet Radio Techniques
2
1.3
Interpolation
4
Comparison of Mobile Telephone Schemes
1.4
6
Further Applications
1.5
7
2.
A Digital Mobile Telephone Scheme
ALOHA Schemes
2.1
8
Digital Encoding of Speech
Digital Transmission in the Urban Environment
A Basic Mobile Telephone System
2.2
15
2.3
21
2.4
30
32
Design of a TDM System
Conclusions
2.5
38
2.6
Interpolation of Speech in Packets
3.
Interpolation
40
3.1
41
Speech Characteristics
3.2
53
The Request Channel
3.3
Delays in the Digital Schemes
60
3.4
Subpackets
Subjective Effects of Telephone Delays
3.5
65
69
3.6
73
3.7
Conclusions
(iii)
Glitches in Packet Speech Interpolation
Slot Assignment Delay
Theoretical Glitch Probability
Glitch Rate and the Use of Subpackets
Computer Simulation Programme
Optimization of Subpackets
Conclusions
4.
4.1
76
4.2
78
4.3
84
4.4
89
4.5
97
4.6
105
The Complete Mobile Telephone System
System Simulation
Effect of the Request Channel
5.
5.1
107
5.2
112
Dependence of Glitch Rate on System Parameters
Optimization of System Parameters
5.3
121
5.4
128
Distribution of Glitches
5.5
136
Subjective Speech Trials
5.6
147
Conclusions
5.7
153
A Comparison of Mobile Telephone Schemes
The Complete Packet Interpolation System
6.
6.1
156
Alternative Mobile Telephone Systems
6.2
162
Comparison of Digital and Small Cell Systems
6.3
165
Small Cell Packet Systems
171
6.4
Conclusions
175
6.5
A Pure TASI Mobile Telephone System
System Description
Theoretical Freezeout Fraction
7.
178
7.1
179
7.2
186
7.3
Effects of the Request Channel
Simulation of the TASI System
189
7.4
191
Comparison with the Packet System
7.5
(iv)
TASI System with Delay
7.6
197
Conclusions
7.7
202
PCM and Other Alternative Applications
TASI-PCM Systems
Packet PCM Systems
Alternative Digital Speech Interpolation Methods
Interpolation in Satellite Channels
Integrated Speech and Data Networks
8.
8.1
205
8.2
210
8.3
215
217
8.4
219
8.5
221
Conclusions
8.6
Conclusions and Further Work
224
9.
Determination of the correct queue arrangement
Queues with dependent inputs
Appendix A
Appendix B
Appendix C
228
231
Simulation programmes and flowcharts
235
250
References
I
1
\
(V)
ABSTRACT
Mobile telephone services are presently provided by frequency
A voice circuit in the form of a
division multiplexing techniques.
radio frequency channel is allocated to a subscriber for his sole use
during a telephone call,
division multiplexing and provide voice circuits on a digital basis.
An alternative to this is to employ time
In such a scheme speech is digitized and formed into packets
for transmission. It is possible to interpolate packets from different
conversations onto a single high capacity radio channel. This is a
form of time division multiple access and is best implemented by a
technique termed reservation ALOHA. When a packet is prepared at a
mobile telephone a request is sent, over a small capacity channel, for
transmission space in the main or speech channel.
To determine the efficiency of this process the characteristics
of speech in packets are investigated through computer simulation. An
optimum packet length is shown to exist and is obtained for a wide
variety of different systems.
Various delays are inherent in packet speech interpolation and
The most important is the delay involved
these are considered in turn.
in assigning space in the speech channel in times of greater than aver¬
age speech activity. Limiting this delay to maintain reasonable speech
quality results in certain packets not being transmitted. Such packets
are said to have been glitched.
The nature and extent of glitches is studied both theoretically
and by further simulation. Their subjective effects are also considered
(vi)
and acceptable rates of speech loss are determined,
straint and another imposed upon the total delay, an optimized packet
This is capable of servicing over 4000 subscribers
Within this con-
system is designed,
in a total capacity of 2.16 Mbit/s.
A comparison is performed between the above packet system and
alternative small cell frequency modulated (FM) systems which employ
It is shown that with similar channel repetition
the packet scheme can provide more telephone connections within a given
There are however penalties of greater cost and complexity.
A compromise system incorporating interpolation with the small cell FM
structure is shown to be superior to both, at least for applications in
the near future.
radio channel reuse.
bandwidth.
Alternative areas for packet speech interpolation such as pulse
code modulation telephone links, satellite systems and integrated data
and speech networks are also considered.
much parallel the mobile telephone situation and that the results
obtained for the latter are of direct relevance.
It is found that these very
In some ways these
areas are better suited to the use of packet speech interpolation than
is a mobile telephone system and they offer the best opportunity for its
implementation at the present time.

								
To top