Green Software Defined Radios: Enabling seamless connectivity while saving on hardware and energy

Series on Integrated Circuits and Systems

Liesbet Van der Perre Antoine Dejonghe

•

Jan Craninckx

Green Software Defined Radios Enabling seamless connectivity while saving on hardware and energy

ABC

Liesbet Van der Perre IMEC VZW Kapeldreef 75 3001 Leuven Heverlee Belgium [email protected]

Antoine Dejonghe IMEC VZW Kapeldreef 75 3001 Leuven Heverlee Belgium [email protected]

Jan Craninckx IMEC VZW Kapeldreef 75 3001 Leuven Heverlee Belgium Jan.Craninckx@ imec.be

ISBN 978-1-4020-8210-8

e-ISBN 978-1-4020-8212-2

Library of Congress Control Number: 2008937503 c 2009 Springer Science+Business Media B.V. ° All Rights Reserved No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper 9 8 7 6 5 4 3 2 1 springer.com

Preface

Green Software Defined Radios, the title of this book may have originated from a lack of inspiration, and the combination of hard work, jet lag, and drinking green tea. The message we want to convey however, is that SDRs are a promising technology for the future, providing they are designed for efficient usage of scarce resources: energy and spectrum. In the last years, the R&D teams focusing on wireless communication (around the world and at IMEC specifically), have realized great breakthroughs. It is our honor, building on this knowledge, to bring a comprehensive overview of the essential technologies. We are grateful that Springer is willing to publish in their collection on radio technologies, a book on green SDRs, a weird species still today, yet maybe the baseline for the day after tomorrow. Dear reader, we wish that you find in the following pages, including the references, some interesting insights, and that this book may live more or less up to your expectations (and hopefully more than less). This book’s closing states that the quest for Green SDRs has not ended, this is just the beginning. Concerning this book however, we are happy that today the opposite is true. We want to acknowledge our colleagues at IMEC for their great scientific contribution, and even more for the enjoyable cooperation. A¨ıssa, Amir, Andr´e, Andy, Andy, Bjorn, Bjorn, Bruno, Boris, Carolina, Charlotte, Claude, David, Dries, Eduardo, Erik, Filip, Frederik, Geert, Gert, Gregory, Hans, Jeroen, Jonathan, Joris, Julien, Lieven, Luc, Maciej, Mark, Martin, Michael, Michael, Michael, Miguel, Min, Mingxu, Noman, Osman, Peter, Pierluigi, Piet, Rodolfo, Roeland, S´ebastien, Sofie, Stefaan, Steven, Thierry, Thomas, Tom, Tom, Tong, Val´ery, Veerle, Vito: hartelijk dank! Leuven, Belgium July 2008

Liesbet Van der Perre Jan Craninckx Antoine Dejonghe

v

Contents

1

2

The Wireless and Technology Scene: Trends Asking for Green Software Defined Radio Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Chronicle of an Innovative Encounter: When Wireless Communication and Micro-electronics Meet . . . . . . . . . . . . . . . . . . . . 1.1.1 The Pioneers’ Era . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.2 The Digital Revolution: When Wireless Communication and Microelectronics Meet . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.3 A Bright Future Ahead? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 The Wireless Scene: Heterogeneity Desires Flexibility . . . . . . . . . . . 1.2.1 Wireless Standards: The Variety is Large, and Growing . . . . 1.2.2 Wireless Terminals go Multi-mode: A Market Perspective . . 1.2.3 Multi-mode Handsets: Enabling Seamless Connectivity . . . . 1.3 The Technology Scene: Cost Imposes Reconfigurability . . . . . . . . . . 1.3.1 Scaling Pleads for Multi-purpose Devices . . . . . . . . . . . . . . . . 1.3.2 Multi-mode Terminals Ask for Hardware Reuse . . . . . . . . . . 1.4 Uniting Wireless Wishes with Technological Constraints: The Power and Spectral Challenge . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.1 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.2 Towards Green Software Defined Radios . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Software Defined Radios: Enabling Seamless Connectivity for Handheld Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Flexible Radios: Species and their Territorium . . . . . . . . . . . . . . . . . . 2.1.1 Flexibility in the Wireless World . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Ancestors: Dedicated Radios . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3 Software Radios: A Designer’s Ultimate Nightmare . . . . . . . 2.1.4 Software Defined Radios: Addressing the Dilemma . . . . . . . 2.1.5 A Debatable Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.6 SW: Brains for SDRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.7 Adaptive Radios: How they (Do Not) Behave . . . . . . . . . . . . .

1 1 1 2 2 3 3 4 5 8 8 10 11 11 12 13 15 15 15 16 17 17 18 18 19

vii

viii

Contents

2.1.8 2.1.9

Multi-modal/Multi-standard Terminals . . . . . . . . . . . . . . . . . . From Flexible Radio to Seamless Services: Standardization Initiatives Paving the Way . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Towards Green SDRs: A Holistic Approach . . . . . . . . . . . . . . . . . . . . 2.2.1 Low Power: A Philosophy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Wireless Communication Scenes: Dynamics are Everywhere 2.2.3 SDR Solutions: Scalability should be Everywhere . . . . . . . . . 2.2.4 Exploit Dynamics and Scalability! . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20 21 22 22 23 24 24 25

3

Software-Defined Radio Front-Ends: Scalable Waves in the Air . . . . . . 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 System-Level Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Wideband LO Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 3–5 GHz Voltage-Controlled Oscillator . . . . . . . . . . . . . . . . . . 3.3.2 0.1–6 GHz Quadrature Generation . . . . . . . . . . . . . . . . . . . . . . 3.4 Receiver Building Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1 MEMS-Enabled Dual-Band Low-Noise Amplifier . . . . . . . . . 3.4.2 Wideband Low-Noise Amplifiers . . . . . . . . . . . . . . . . . . . . . . . 3.4.3 Wideband Downconversion Mixer . . . . . . . . . . . . . . . . . . . . . . 3.4.4 Flexible Baseband Analog Circuits . . . . . . . . . . . . . . . . . . . . . 3.4.5 Analog-to-Digital Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Transmitter Building Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Calibration Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.1 Quadrature Imbalance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.2 DC-Offset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.3 Impact of LPF Spectral Behavior . . . . . . . . . . . . . . . . . . . . . . . 3.7 Full SDR Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27 27 28 30 30 37 39 40 41 45 46 51 54 57 57 58 58 59 61 62

4

SDR Baseband Platforms: Opportunism to Combine Flexibility and Low Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 SDR Baseband Platforms: Going Mobile . . . . . . . . . . . . . . . . . . . . . . . 4.2 Approach to Combine Flexibility and Low Energy: Divide and Conquer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1 Opportunistic Partitioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 Low Power Operation: Sleeping, Waking, and Working on Minimal Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.3 SDR Platform Implementation: Teaming up with Deep-Submicron Technology . . . . . . . . . . . . . . . . . . . . . . 4.3 Digital Front-End: Going Reactive and Cognitive . . . . . . . . . . . . . . . . 4.3.1 The Global DFE: Speaking and Listening Means for the Baseband Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.2 Zooming in on the Power Detection and AGC Controller . . . 4.3.3 Zooming in on the Synchronization Engine . . . . . . . . . . . . . .

65 65 66 66 68 70 74 74 76 76

Contents

Processors for SDR-Baseband: Working Horses in a Race for Speed and Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1 The Quest for High Performance and Low Power: Introducing Different Styles . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2 Tuning an ADRES Processor: A Suitable Case . . . . . . . . . . . 4.5 Outer Modem Engine: Going with the Flexibility Stream . . . . . . . . . 4.5.1 Problems with Dedicated Solutions Arising . . . . . . . . . . . . . . 4.5.2 Flexible Solutions in Sight . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.1 SDR Baseband Platforms: Going Mobile Today . . . . . . . . . . . 4.6.2 The Future: Next Generations Desired . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

4.4

80 80 81 88 88 89 93 93 93 94

5

Software: Fuel for Green Radios: The Blessing and the Curse . . . . . . . . 97 5.1 The Blessing and the Curse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 5.2 Structured SW Design: Going for Network and Platform Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 5.3 Platform-Level SW: The Control Room for the SDR . . . . . . . . . . . . . 100 5.3.1 The Strategic Plan: Design Flow . . . . . . . . . . . . . . . . . . . . . . . 100 5.3.2 Meeting the Design Goals: Latency Requirements . . . . . . . . . 100 5.4 Baseband Processor SW: The Working Horse for the SDR . . . . . . . . 101 5.4.1 The Strategic Plan: Design Flow . . . . . . . . . . . . . . . . . . . . . . . 101 5.4.2 Meeting the Design Goals: Real-Time Requirements . . . . . . 103 5.5 System Level SW: Providing SDR Terminals with Social Skills . . . . 105 5.5.1 The Strategic Plan: A Simulation Framework for Network Centric SW Development and Validation . . . . . . . . . . . . . . . . 105 5.5.2 Meeting the Design Goals: The 802.11n Case . . . . . . . . . . . . 107 5.6 Future Challenges and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 5.6.1 The Wireless Race for More: Trouble Ahead . . . . . . . . . . . . . 109 5.6.2 Solutions to Boost Performance: More Parallelism in the SW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 5.6.3 Solutions to Save Power: Architecture-Aware Scalable SW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 5.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

6

Energy-Aware Cross-Layer Radio Management: Exploit Flexibility for Saving Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 6.2 SDR Design Step: Enable Flexibility and Energy Scalability . . . . . . 118 6.2.1 Reconfigurable Analog Front-End . . . . . . . . . . . . . . . . . . . . . . 118 6.2.2 SDR Digital Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 6.3 SDR Control Step: Exploit Flexibility and Scalability for Saving Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 6.3.1 State-of-the-Art Energy Management Techniques . . . . . . . . . 122

x

Contents

6.3.2 Cross-Layer Performance-Energy Optimization . . . . . . . . . . . 123 6.3.3 Instantiation in a Use Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 7

Towards Cognitive Radios: Getting the Best Out of the Radio and the Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 7.1.1 The Need for Reconfigurable Radio Platforms . . . . . . . . . . . . 135 7.1.2 The Need for Intelligent and Adaptive Radio . . . . . . . . . . . . . 136 7.2 New Control Functionality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 7.2.1 Cognitive Radio: Broad View . . . . . . . . . . . . . . . . . . . . . . . . . . 138 7.2.2 Cognitive Radio: Spectrum-centric View . . . . . . . . . . . . . . . . . 140 7.3 New Sensing Functionality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 7.4 New Radio Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 7.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

8

Close: This is not the End, it’s Just a Beginning . . . . . . . . . . . . . . . . . . . . 153 8.1 A Last Chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 8.2 Major Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 8.3 Challenges Ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 8.3.1 Scaling to Next Generation Applications and Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 8.3.2 Focus on Multi-band Antenna Interface Challenge . . . . . . . . 154 8.4 Closing Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Chapter 1

The Wireless and Technology Scene Trends Asking for Green Software Defined Radio Solutions

1.1 Chronicle of an Innovative Encounter: When Wireless Communication and Micro-electronics Meet A radio is defined as ‘a communication system employing wireless transmission of information by means of electromagnetic waves propagated through space’. Radio became reality thanks to brilliant theorists and creative experimentalists. Yet, the encounter with micro-electronics brought the impressive boost we witness today, and plays a crucial role in the evolution towards Software Defined Radios. In the next paragraphs, a brief history rolling into the future is outlined.

1.1.1 The Pioneers’ Era James Clerk Maxwell in 1864 wrote his paper ‘A Dynamical Theory of the Electromagnetic Field’ and derived the famous equations named after him. Little could he know what would be the future impact of his findings! In 1887, Heinrich Hertz was able to confirm Maxwell’s thoughts with impressive successful experiments: the propagation of electromagnetic waves in free space was proven. This milestone launched the quest for possibilities to ‘stick’ information on radio waves, and thus ‘tele-communicate’ without the need to install wires and be restricted by them. Guglielmo Marconi was the first to obtain a patent in the field of wireless communication in 1896, the year often mentioned for the invention of radio, and he later received also a Nobel Prize for his work. The first transmission of voice and music via radio-waves, on Christmas evening 1906 by Reginald Fessenden, gave birth to radio broadcasting. Importantly, this breakthrough offered the technology a human voice and face, raising its interest to a very large public.

L. Van der Perre et al., Green Software Defined Radios, Series on Integrated Circuits and Systems, c Springer Science+Business Media B.V. 2009 

1

2

1 The Wireless and Technology Scene

1.1.2 The Digital Revolution: When Wireless Communication and Microelectronics Meet A dramatic increase and improvement of wireless communication took place since, towards the current so-called ‘third generation’ (3G) [1] and WiFi [2] multimedia communication devices that allow to exchange voice, image and data information at speeds of respectively a couple to over 50 megabits per second (Mb/s). This evolution in communications has been particularly accelerated in the second half of the twentieth century thanks to the independent evolution of two fields of science: information theory and microelectronics. The year 1948 is considered as the major landmark in the development of digital technology due to achievements in both fields that would be eventually successfully combined. Claude Shannon on the information theory side set the very founding stone of this science with the definition of one binary unit of information, or bit, as well as the definition of channel capacity. In the same year, William Shockley and his team at Bell laboratories announced the invention of the transistor, which would be used later as the building element of circuits able to process and store bits.1

1.1.3 A Bright Future Ahead? Today, the combination of mobility and connectivity has become a commodity and an essential comfort in today’s society. Remarkably, even 70 year old people declare not to be able to miss this technology anymore, while they lived most of their life without it. The variety of wireless standards, supporting a large range of services in different geographical environments and regions, is large and growing. Technology is become ever smaller and faster. These two evolutions, explained in the following sections, together ask for Software Defined Radio (SDR) solutions. The gigantic success of wireless communication is however forming a threat to itself. The analogy can be made: the convenience and popularity of car transportation has lead to enormous traffic jams hampering mobility, and unsustainable energy consumption. Similarly, the spectrum gets crowded, and increased service requirements are draining mobile devices’ batteries.

1

This parallel between Shannon’s definition of a bit and the transistor was based on Berrou’s and Glavieux’s introduction to their IEEE Information Theory invited paper of 1998, on the occasion of the 50th birthday of the transistor and of information theory. The whole article can be found in http://www.itsoc.org/review/frrev.html.

1.2 The Wireless Scene: Heterogeneity Desires Flexibility

3

1.2 The Wireless Scene: Heterogeneity Desires Flexibility 1.2.1 Wireless Standards: The Variety is Large, and Growing Wireless communications are routinely used today for a large variety of applications including voice, data transfer, Internet access, audio and video streaming, just to name a few. For one specific service, several systems are standardized, and they can each become the preferred option in several regions of the world. Table 1.1 illustrates this for digital broadcasting systems. Just to support this service on a terminal, already some degree of flexibility in the radios is needed. Pushed by the insatiable demand for bandwidth and pulled by the steady improvement of semiconductor technology (Moore’s law), the performance offered by wireless standards is due to improve with the years, seemingly without bounds. This is illustrated in Fig. 1.1 which shows the observed and predicted evolution of the main classes of wireless access standards as thick arrows (from bottom to top): Wireless Personal Area Networks (WPAN), Wireless Local Area Networks (WLAN), Wireless Metropolitan Area Networks (WMAN), and Wireless Wide Area or cellular (WWAN) networks. Obviously, higher rates will be more easily – or sooner – achieved at low mobility than at higher mobility. Table 1.1 Various standards for digital video broadcasting

4

1 The Wireless and Technology Scene 1995

High speed

2000

2G (digital)

2005

2010

3G+

3G Multimedia

GPRS UMTS EDGE CDMA2000 GSM CDMAone

3GPPLTE+

802.16e

4G

Medium speed

research target

1G (analog) WIMAX

Low speed/ Stationary

2.4 GHz WLAN

5 GHz WLAN

UWB WPAN

Bluetooth

10 kbps 100 kbps 1 Mbps

High rate WLAN

10 Mbps

100 Mbps

60 GHz WPAN

1 Gbps

Fig. 1.1 Variety of wireless access standards

1.2.2 Wireless Terminals go Multi-mode: A Market Perspective From a user perspective, it is very attractive to have a single handheld device that can support a large variety of wireless standards. In an answer to the user’s demand, mobile handsets have started supporting multiple modes over the past years. In a first instance, this is achieved by integrating multiple radios into one handset. A nice example is shown in Fig. 1.2 [2]. Clearly, when the number of radios increases, the cost, size, and weight of the terminal are seriously affected by the multi mode extension. A much more efficient solution is clearly offered by radios which you can reconfigure (such as e.g. Software Defined Radios) to access several wireless standards. This is further explained in Section 3.2. Market forecasts [1] predict that the partition of ‘SDR enabled’ mobile handset shipments will grow considerably in the coming years, really ‘taking off’ probably in 2010 (see Fig. 1.3). Importantly, going from dedicated radios to reconfigurable radios, brings about a real paradigm shift! Consequently, manufacturers only take the leap when significant advantage is expected. Maybe they should even ‘feel the pain’ first. In NorthAmerica, more dispersion in cellular standard adoption occurs, with both CDMA and GSM systems being widely used. As a result, the proportion of SDR enabled mobile shipments, is increasing much faster in this region, as shown in Fig. 1.4 [1]. For all clarity and from the point of completeness, the global distribution of absolute numbers (optimistic case) SDR enabled handset shipments is given in Fig. 1.5 [1]. Clearly, as for dedicated radios, the largest market in terms of volume is situated in the Asia Pacific region.

1.2 The Wireless Scene: Heterogeneity Desires Flexibility 1

2

3

RF ID

FM

5 4 5

6

7

Antennas

UWB

WLAN

Blue tooth

8

2G/3G Cellular

DVB-H

9

10

diversity RX

GPS

11

Fig. 1.2 Multi-mode handset featuring separate radios (Nokia [2])

12

Optimistic

Pessimistic

% of total shipments

10

8

“SDR enabled” = handset with programmable baseband processor

6

4

2

0 2007

2008

2009

2010

2011

Fig. 1.3 SDR enabled handset shipments 2007–2011 worldwide [1]

1.2.3 Multi-mode Handsets: Enabling Seamless Connectivity Ubiquitous and seamless connectivity can be achieved in a heterogeneous network environment, under the condition that both terminals and network enable feature the necessary reconfiguration capabilities to support horizontal (between access point adhering to one standard) and vertical (between access points operating different standards) roaming. Recently, the need for reconfiguration support is receiving

6

1 The Wireless and Technology Scene 30

Optimistic

Pessimistic

% of total shipments

25

20

15

10

5

0 2007

2008

2009

2010

2011

Fig. 1.4 SDR enabled handset shipment in North-America 2007–2011 ([1])

100%

80%

60%

40%

20%

0% 2009 - Optimistic

2010 - Optimistic

2011 - Optimistic

Fig. 1.5 Global breakdown of SDR enabled mobile shipment 2007–2011 [1]

attention in specific standardization initiatives, as illustrated in Fig. 1.6 (from IEEE SCC41 [4], formerly IEEE 1900). This confirms that technological answers are due to answer the user’s need for seamless connectivity. These technological answers need to include:

1.2 The Wireless Scene: Heterogeneity Desires Flexibility

7

Fig. 1.6 Need for reconfiguration support to enable ubiquitous and seamless connectivity

Multi-mode radios: user terminals should encompass radios that can generate the different waveforms as specified in the various standards. One solution to generate these various waveforms is to conceive flexible radios. In Chapter 2 we will position Software Defined Radios as a solution to realize low power flexible radios. Control solutions for reconfiguration: Control solutions for flexible radios should assure reliable connectivity, allocating the available resources in the most efficient way. In this book, we express that this challenge should be considered and optimized across the different OSI-layer, and we propose an approach to realize cross-layer (XL) control solutions. On top of to seamless connectivity to one service, users want to enjoy a multitude of services on one terminal: Streaming interactive connectivity (voice and video), high-speed data access, broadcasting reception, and short range connectivity to devices in the close proximity. Predictions claim that handheld devices will need to support at least six different radios already in 2009. In conclusion from a functionality point of view, flexibility is desired! A closer inspection of the communication schemes (see Table 1.2 for a summary of the major specifications for key wireless standards) reveals that this concept is quite challenging. Indeed, the variety of bit rates, modulation formats, physical bandwidths and carrier frequencies, is large. Fortunately, we see some common trends in broadband access schemes, which enable to optimize flexibility in the radios. For example, modulation schemes applying frequency domain processing are recurrently used

8

1 The Wireless and Technology Scene

Table 1.2 Major characteristics of wireless access standards

for achieving high rates in fading environments. Also, the use of multiple antennas processing, in its most advanced flavor ‘Multiple Input Multiple Output’ (MIMO), is becoming commonplace. The assessment of and the impact on SDR front-end and baseband platform, will be given in the relevant chapters further in this book.

1.3 The Technology Scene: Cost Imposes Reconfigurability 1.3.1 Scaling Pleads for Multi-purpose Devices In parallel to the increasing data rates and need for functional flexibility in radio systems, progress in CMOS technology also has witnessed an impressive evolution over the last decades. Gordon E. Moore predicted correctly 40 years ago, that the number of transistors on a chip would double about every 2 years. He added “. . . (T)he first microprocessor only had 22 hundred transistors. We are looking at something a million times that complex in the next generations-a billion transistors. What that gives us in the way of flexibility to design products is phenomenal.” This evolution is actualized and detailed in the International Technology Roadmap for Semiconductors (ITRS roadmap) [3]. Indeed, transistors can be made ever smaller

1.3 The Technology Scene: Cost Imposes Reconfigurability Fig. 1.7 CMOS scaling makes chips ever faster

9 1400 GHz

ITRS 2003 fT

150 GHz

45 nm 18 nm

100 nm

$1M

Mask cost

$400k

350 nm

250 nm

180 nm

130 nm 90 nm

65 nm

45 nm

Fig. 1.8 Mask costs are increasing exponentially

and faster (see Fig. 1.7). The scaling has brought enormous processing capabilities in small areas, opening opportunities to implement flexible platforms at low cost and low power. For the last decades, the ‘happy scaling’ has offered us more functionality and at the same time lower costs. Yet, for the newer technology nodes, the Non-RecurringEngineering (NRE) costs related to system-on-chip (SoC) design are rising exponentially. This is illustrated for the mask cost in Fig. 1.8. Next to the mask cost, the design cost has increased dramatically, and is expected to do so. Not only is the complexity of the designs increasing, moreover CMOS scaling has arrived at the point where parasitic problems are becoming dominant: variability, reliability, and last but not least leakage. These effects can not be resolved anymore at the technology (transistor) only, and have to be taken care off in the design phase as far as possible. The ‘problem table’ in Table 1.3 indicates which parasitic effects are expected to impact analog and digital design for 45 nm and smaller technologies. Today, the question is often asked whether for cost reasons, scaling is still preferred. Till now however, it seems that the rising NRE costs have been compensated by the fact that ever higher volumes are produced. For mass markets, such as for

10

1 The Wireless and Technology Scene

Table 1.3 CMOS scaling parasitic problems for analog and digital design Problems for 45 nm and beyond Increase of gate leakage current Increase of subthreshold leakage currents Increase of variability Decrease of voltage gain Degradation 1/f noise by high -k gate dielectrics Increase of switching noise Increase of Miller capacitance Increase of transistor series resistance VDD decrease reliability Inductors consume much expensive area Fig. 1.9 Scaling brings cost advantage for high-volume products

For digital

For analog RF

X X X

X X X X X X X X X X

X X X X

$/chip

Not you!

Node x Node x+1

volume

mobile devices, scaling from technology node ‘x’ to the next one ‘x + 1’ is still a must to be competitive (see Fig. 1.9). Of course, this pleads for multi-purpose devices. Already today in specific cases, cost trade-offs show a cost advantage in using a reconfigurable radio for singlemode devices as well: the extra area penalty is there not significant compared to the NRE. ‘Flexible radio extremists’ will even claim that on the (extreme?) longer term, no dedicated radios will survive.

1.3.2 Multi-mode Terminals Ask for Hardware Reuse For multi-mode terminals, clearly the possibility to reuse silicon and thus significantly reduce overall area makes the argument to go for flexible platforms much stronger still. Moreover, other cost factors also direct towards hardware reuse: Number of components and assembly cost: Some visions claim that ‘all the cost will be in the assembly’. While current terminal breakdowns do not show this yet, intuitively indeed we can foresee this share of the cost will grow in the future.

1.4 Wireless Wishes with Technological Constraints: The Power and Spectral Challenge

11

Replacing multiple radios by a single flexible radio, clearly reduces the number of components and consequently the assembly cost. Form factor (size and weight): The chase for ever flatter and lighter mobile terminals is constant, while more and more features and radio interfaces need to be supported. Multi-purpose devices bring an important, in the future maybe even indispensable, advantage. Time to market: New standards, updates, regional flavors, seem to pop up every day, while designing a full new radio easily takes years. The time to market can be decreased impressively, if redesigning can be traded for reconfiguring and reprogramming. The effort hidden in the reprogramming should however definitely not be underestimated (see ‘blessing and curse’ in Chapter 5). Taking into consideration the scaling evolution and constraints for multi-mode radios, we can conclude that cost imposes reconfigurability, and will drive manufacturers towards flexible radios for future wireless terminals.

1.4 Uniting Wireless Wishes with Technological Constraints: The Power and Spectral Challenge 1.4.1 Challenges The combination of the increasing need for functional flexibility and the huge cost related to SoC design will make implementation of multi-mode radios on SDR platforms the only viable option in the future. Mobile devices being battery-powered, the performance requirements are coupled with severe constraints on energy efficiency. This is becoming a key concern: there exists a continuously growing gap between the available energy, resulting from battery technology evolution, and the steeply increasing energy requirements of emerging radio systems (see Fig. 1.10). A major challenge therefore is to enablelow energy reconfigurable radio implementations, suited for low-cost handheld multimedia terminals. They should reach battery life-time of today’s fixed hardware implementations, and at the same time offer reliable connectivity. Besides the energy constraint, spectrum is also becoming a major resource bottleneck. The spectrum is a scarce valuable resource, which has become over-allocated over the past decades. Figure 1.11 shows a sample of the allocation of the spectrum below 3 GHz, clearly illustrating the congestion. Due to the accelerated deployment of broadband personal communication and the continuously increasing demand for higher data rates, we are heading towards a red brick wall. New paradigms for efficiently exploiting the spectrum are obviously needed. A current trend is the evolution towards dynamic and open access to spectrum, motivated by the under-utilization of many licensed frequency bands. This has led to

12

1 The Wireless and Technology Scene

Fig. 1.10 The energy gap between required and available energy is growing Gap Energy requirement

Energy available in battery

Time

Fig. 1.11 Spectrum allocation snapshot: no space left

the concept of cognitive radio (CR, see Chapter 7 for a clarification of nomenclature). Cognitive radios will essentially need flexibility support in HW and SW and adequate control for reconfigurability.

1.4.2 Towards Green Software Defined Radios In this book, we present a holistic approach towards ‘Green Software Defined Radios’, enabling seamless connectivity while saving on hardware and energy. In Chapter 2, the technical content of the book is introduced. Relevant taxonomy is given and Software Defined Radios are positioned in the broader sphere of flexible radios. A holistic system approach towards low cost, low energy SDRs is presented, leveraging on the concept of providing and exploiting scalability for low energy. Radio front-end building blocks need to be designed to offer flexibility in carrier frequency, channel bandwidth, noise performance, etc. without a significant power penalty. An overview of SDR front-end challenges and solutions is given in Chapter 3. We specifically also zoom in to a scalable ADC architecture, that nicely couples scalability to a low power consumption. In Chapter 4 a survey of SDR baseband solutions is given. Special focus is on a heterogeneous Multi-Processor System-on-Chip (MPSoC) approach optimized for scalability and low energy. Particular attention is paid to the optimization of domainspecific processors optimized for wireless communications functionality.

References

13

On top of the hardware aspects, software is obviously of paramount importance when designing Software-Defined radios. In Chapter 5 solutions and approaches are presented to enable efficient design both for platform-level and processor-specific SW components. Chapter 6 addresses the topic of cross layer optimization, resulting in control solutions which exploit flexibility for low energy. A generic approach to master optimization complexity will be given. Moreover, use cases will be presented showing the impressive gains that can be achieved. The spectral challenge is the focus of Chapter 7. A preview is given on how SDRs are crucial to realize cognitive radios, and it is explained which are the specific features that will need to be added. Chapter 8 closes the book, summarizing the major contributions. Also, some open research questions are put forwards. Specifically evolutions in both the wireless and the technology are considered. Importantly we remark that, given the aggravating energy, capacity, and flexibility requirements, the quest for green SDRs will not end in the foreseeable future.

References 1. 2. 3. 4.

Software Defined Radio in Mobile Phones, ARCchart, November 2007 Yrj¨o Neuvo, CTO, Nokia at ISSCC2004 www.itrs.net www.scc41.org

Chapter 2

Software Defined Radios Enabling Seamless Connectivity for Handheld Devices

2.1 Flexible Radios: Species and their Territorium A continuous trend towards ‘more flexibility’ in wireless systems is witnessed. Before digging into the topic of Software Defined Radios, we position them in the perspective of this trend. First, flexibility in the overall wireless context is introduced. Next a (view on) taxonomy of flexible radios is given. An analysis of the software in SDRs is given. Finally, some relevant standardization initiatives are listed.

2.1.1 Flexibility in the Wireless World Flexibility is a property of radio systems and networks, and thus it does not get easily categorized within the various layers of the familiar OSI communication model (or the related IP network model). Rather, it manifests itself across the layers, and its scientific inquiry calls for a multi-layered approach, one that explores synergies rather than separations. One can distinguish between platform (or equipment)-centric flexibility, network-centric flexibility and service/applicationcentric flexibility. The network-centric flexibility addresses radio-network-architecture features, where the flexibility (potentially including reconfigurability) aspects of the platform-centric portion can be abstracted. The Service/Application-centric flexibility is more and more present in wireless services and applications. This flexibility can be nicely coupled to the radio flexibility, and synergies can be exploited (see Chapter 6). The flexible radios discussed in this book, essentially generate flexible waveforms to be sent over the air. This is achieved through reconfigurability in the wireless modem(s). Consequently, they concern relevant functionalities of the HW and L. Van der Perre et al., Green Software Defined Radios, Series on Integrated Circuits and Systems, c Springer Science+Business Media B.V. 2009 

15

16

2 Software Defined Radios

the SW, radio- (RF) and intermediate frequency (IF) front-ends, digital Base-Band (BB) platforms, and control strategies. For the latter, a cross-layer approach is proposed to achieve the best power/performance accords. The focus is on handheld terminals, which essentially should live on (low weight) batteries, and can feature a high mobility and end up in a large variety of geographical environments. In the coming sections we shortly introduce taxonomy of flexible radios, in order to position the specific species of Software Defined Radios which form the subject of this book. This taxonomy has grown to some extent organically, and therefore is not perfectly unambiguous. The interpretation proposed in this book builds upon the classification proposed by communities of experts,1 and the authors’ personal preference.

2.1.2 Ancestors: Dedicated Radios Traditionally, radios have featured little or no flexibility, in the sense that they have been designed to be compatible to a specific standard. We call this class dedicated radios. More recently, a certain degree of flexibility in the waveform has been introduced within one standard. A clear example is the recent IEEE802.11n, where not only different constellation sizes and code rates should be supported, but also several operating frequencies and bandwidths have been defined [12]. The emerging 3GPPLTE standard [13] is envisaged to open even more flexibility opportunities to fit in a wide variety of wireless communication scenes (see definition in Section 2.2). These standards clearly require the radio to handle different parameter settings. Yet, they have till now still mostly been built for one purpose (standard and/or service), and therefore receive the ‘dedicated’ label. Two flavors of non-flexible radios are categorized: A Hardware Radio (HR) is implemented using hardware components only and cannot be modified except through physical intervention. This radio is becoming an endangered species, and even extremely low power devices more and more ask for some scalability [6]. A Software Controlled Radio (SCR) implements only the control functions in software – thus only limited functions are changeable using software. This can for example be used to control parameters of a modem. An exemplary OFDM-modem for WLAN is documented in [3].

1 Specifically, the SDR forum (www.sdrforum.org) definitions and terminology used in the WWRF forum 0 have been used as a reference. Private communication with Prof. Andreas Polydoros and discussion in the context of the NEWCOM project have also contributed significantly to throw light in the matter.

2.1 Flexible Radios: Species and their Territorium

17

2.1.3 Software Radios: A Designer’s Ultimate Nightmare One of the most far-fetched ways to implement flexible radios, is to apply the concept of Software Radio (SR) [2] that basically assumes that all the radio modules are implemented by software. These radios seem to offer the ultimate freedom many aspire [9]. The architecture is quite simple, as illustrated in Fig. 2.1: place the Analog-to-Digital Convertor (ADC) and Digital-to-Analog convertor (DAC) at the antenna, and perform all transmission and reception functionality digitally, preferably on a DSP. The clear advantage of this kind of radios is that they can easily be modified to adjust to new services and management strategies. Hence, they are each marketing manager’s dream. Yet, for the RF IC designer they are a nightmare come true. Considering the specifications of the ADC, one would need a sampling rate in the order of 10 GS/s, at a resolution of 16 bits. The power consumption of the ADC alone would in the order of 100 W, assuming only you could ever design it. This clearly does not fit in the power budget of a mobile device. As for the further processing of these samples coming in at 10 GS/s at 16 bit on a DSP, this is also a job one would not like to see happening on a handheld device.

2.1.4 Software Defined Radios: Addressing the Dilemma A Software Defined Radio (SDR) can be defined [1, 4, 8] as a collection of hardware and software technologies that enable reconfigurable system architectures for wireless networks and user terminals. SDR provides an efficient and comparatively inexpensive solution to the problem of building multi-mode, multi-band, multi-functional wireless devices that can be enhanced using software upgrades. SDR-enabled devices (e.g., handhelds) and equipment (e.g., wireless network infrastructure) can be dynamically programmed in software to reconfigure the characteristics of equipment. This demarcation could be considered as the definition of the broader class of reconfigurable radios. For SDRs specifically, we assume at least a significant part of the functionality is actually implemented in software running on processors, which can be more or less tuned for the domain or sub-functionality

ADC DSP

Fig. 2.1 Software radio concept

DAC

18

2 Software Defined Radios

(see Chapter 4). Reconfigurable radios purely composed of parametrizable HW blocks, can be built to achieve lower power consumption, at the price of offering less flexibility. SDRs thus make use of a common hardware platform, which typically consists of an analogue front-end and a digital baseband platform, incorporating processor(s) to perform part of the data transmission and reception functionality. In the partitioning, a trade-off between degree of flexibility and implementation complexity (and power) is made. Average power consumption and energy/bit are considered as the relevant metrics for which radios for handheld devices should be optimized. In this view, SDRs are in this book advocated as the best answer in the current state-of-theart to provide flexible radios for mobile devices.

2.1.5 A Debatable Overview For the purpose of making the working assumptions of the authors of this book explicit, and at the risk of controversy, an overview of the different species of flexible radios is sketched in Fig. 2.2 below.

2.1.6 SW: Brains for SDRs Reconfigurable HW platforms essentially form the muscles for SDRs. The actual wireless modem functionality is implemented by means of the SW running on the SDR platform and terminal. The SW needs to implement the functionality associated to the different layers in the OSI stack, is implemented on several levels of abstraction, and can run either on the SDR platform or on a host processor in the terminal. Figure 2.3 shows an exemplary build-up of the SW in an SDR system. The platform and processor dependent SW today typically includes all SW implementing the PHY layer, and can include drivers and SW modules which have been highly optimized towards specific processors.

Flexible Radios Reconfigurable Radios Software Defined Radios Software Radios

Fig. 2.2 Species of flexible radios, classification assumed in this book

2.1 Flexible Radios: Species and their Territorium

19

Other protocol functions & management

Embedded SW on SDR Platform

Can include common SW components *

SW on host processor

Platform control & time critical MAC Hardware Abstraction Layer Platform and processor dependent SW

Processor 1

Processor n

‘Operating environment’ ‘firmware’ *

Platform component i

SDR Hardware * Eventually downloadable

Fig. 2.3 Exemplary build-up of the SW in an SDR system

Eventually, new SW could be downloaded on terminals through the network. This definitely is envisioned for platform-independent SW protocol and management functionality. On a level much closer to the platform hardware, one could also foresee ‘downloadable waveforms’ in the future, and early attempts have been made on a simulator level [7]. This basically implies intervening in the firmware of the platform. Operators clearly would enjoy a generic waveform-downloading feature. On the mid term and for handheld devices, power constraints limit the practicality of this prospect. This book primarily focuses on the SDR platform itself, including the platform and processor-dependent SW (see Chapters 4 and 5 for more detail).

2.1.7 Adaptive Radios: How they (Do Not) Behave In the previous sections, different classes of flexible radios have been defined, all featuring the common attribute that they can ‘output’ flexible waveforms. This flexibility relates to the physical possibilities of the radios. Complementary, radios can be categorized on how they will actually behave in operation, ‘at run-time’. In their behaviour, radios can either be adaptive, or not. Figure 2.4 illustrates how the flexibility in the physical radio implementation and the intelligence of the radio (‘Intelligent Signal Processing’ [ISP]) can be considered as two different dimensions. Advances among both dimensions are essential to progress towards ultimate ‘Mitola’ [15] cognitive radios (see Chapter 7).

20

2 Software Defined Radios

flexibility Mitola radio

full SR no limits

full SR limited

SDR

SCR

basic radio

logic

analysis

intuition

ISP

Fig. 2.4 Evolution from basic radio to ‘Mitola’ radio through increasing physical radio flexibility and intelligence

Non-adaptive radios receive a value for all their parameters before starting up a communication, and keep on transmitting and receiving accordingly afterwards. One could say that they ‘do as they’re told’, they ‘behave’. Adaptive radios will not keep the settings they received at initialisation. In stead, they may change some of their parameters based on external conditions, more specifically the dynamics in the wireless communication scene. For example, a WLAN modem lowering its bit rate when the path loss it measures has increased above a certain threshold (e.g. by switching to a lower constellation size), demonstrates adaptive behaviour. There is no one-to-one relationship between flexibility and adaptability. Flexible radios can behave completely non-adaptive, in cases where they are configured before transmission/reception, and consequently stick to the same parameter set. Similarly, dedicated radios can be adaptive, as illustrated in the WLAN example above. Naturally, flexibility in the radio platform opens up wide-ranging adaptation opportunities, crossing boundaries of standards. A specific subset of adaptive radios applying learning techniques, are called cognitive radios. Taxonomy for cognitive radios also has brought about some nice confusion. The relevant terminology is treated in Chapter 7.

2.1.8 Multi-modal/Multi-standard Terminals In the trend towards seamless connectivity in a heterogeneous network environment, multi-modal/multi-standard terminals are gaining momentum on the market [5].

2.1 Flexible Radios: Species and their Territorium

21

Most of the products today comprise parallel chains of dedicated radios for different standards or modes. Recently, reconfigurable radios are entering these terminals, which support a set of modes and/or standards on a common hardware platform. As elaborated on in Chapter 1, cost and size concerns clearly favor reconfigurable radios as the preferred option for multi-modal/multi-standard terminals.

2.1.9 From Flexible Radio to Seamless Services: Standardization Initiatives Paving the Way Innovative research has recently made impressive progress towards flexible radios. In parallel, crucial questions arise on how these radios will be operated in a real network environment, where in the end users should be able to benefit from the interesting new features of these radios. Terminal manufacturers also worry about certification of these devices. Fortunately, standardization initiatives have started concentrating on the aspects of reconfigurability and interoperability.

2.1.9.1 SDRs and their Interoperability One of the expectations towards SDRs is that they will enable the development of ‘common’ or ‘open’ software modules in the future, which could be implemented on various platforms, independent of the actual terminal equipment or manufacturer. In this context, several initiatives focus on standardizing SW modules and interfaces running on the operating environment of an SDR platform, assuming the actual radio platform could be abstracted (in the future). One association active in the definition of essential interfaces in this context, is the Software Defined Radio Forum (SDR forum) [8]. The SDR forum is a nonprofit organization dedicated to promoting the development, deployment and use of software defined radio technologies for advanced wireless systems. As one of the outputs, the technical committee of the forum generates technical reports providing a description of the concepts and basic architecture for an SDR along with a description of the internal interfaces for the radio, its software download process, interfaces between various modules, and basic message definitions needed for such a process. In a security context, the Joint Tactical Radio System (JTRS) [7] is a program of the US and NATO to produce radios which provide flexible and interoperable communications. Examples of radio terminals which require support include handheld, vehicular, airborne and dismounted radios, as well as base-stations (fixed and maritime). This goal is achieved through the use of SDR systems based on an internationally endorsed open Software Communications Architecture (SCA). The Wireless World Research Forum (WWRF) has established a specific Working group on reconfigurability [10]. They also stress the importance of reference models for reconfigurable terminals and network architectures, and have published relevant white papers on the topic.

22

2 Software Defined Radios

2.1.9.2 Towards Dynamic Spectrum Access Networks Dynamic spectrum access can leverage on SDRs, and opens up a new wireless order. While this new paradigm opens up access to unrevealed capacity, along comes a whole collection of new regularization and standardization issues. SCC41 [14] encompasses standards projects in the areas of dynamic spectrum access, cognitive radio, interference management, coordination of wireless systems, advanced spectrum management, and policy languages for next generation radio systems. SCC41 is particularly interested in ideas that could be implemented in commercial products in the near to medium term. The scope of this standardization covers a wide range of aspects related to concepts and technologies in the fields of spectrum management, policy defined radio, adaptive radio, software defined radio, reconfigurable radio and networks: – WG IEEE 1900.1 aims to develop a standard which will facilitate the development of these technologies, by clarifying the terminology and how these technologies relate to each other. – WG IEEE 1900.2 analyzes effects related to coexistence and interference. – WG IEEE 1900.3 focuses on ‘Recommended Practice for Conformance Evaluation of Software Defined Radio (SDR) Software Modules’. The goal of this effort it to assure that SDR software can be deployed with high confidence that it will operate within prescribed regulatory and operational limits. The guideline will apply to wireless network operators and terminal equipment manufacturers to help them define test guidelines that conform to SDR technologies, to be licensed by regulatory authorities. – WG IEEE 1900.4 discusses Architectural Building. The standard defines the building blocks comprising network resource managers, device resource managers, and the information to be exchanged between the building blocks, for enabling coordinated network-device distributed decision making which will aid in the optimization of radio resource usage, including spectrum access control, in heterogeneous wireless access networks (see also Chapter 1). The challenges to realize the necessary standards and regulations for SDRs and cognitive radios are considerable. Yet, relevant initiatives (as described above) are gaining momentum. Major industrial players are teaming up with academia to create the context needed to enable SDRs to be exploited to their full potential.

2.2 Towards Green SDRs: A Holistic Approach 2.2.1 Low Power: A Philosophy Given the energy gap introduced in Chapter 1, a major challenge is to enable low energy reconfigurable radio implementations, suited for handheld multimedia terminals and competitive with fixed hardware implementations. Several major integrated

2.2 Towards Green SDRs: A Holistic Approach

23

device manufacturers have indicated in recent announcements that energy efficiency (i.e., MOPS/mW) is almost not improving any more by technology scaling. The attention for low power can not be localized in one stage of the design: it needs to become a philosophy. A wise team said2 : ‘Think about low power all the time, if it needs an all-time power record!’ Platform improvements and circuit design progress are essential but not sufficient for bridging the energy gap. A clear need for holistic system-level strategies exists, and disruptive solutions are needed.

2.2.2 Wireless Communication Scenes: Dynamics are Everywhere A radio has been defined as a communication system employing wireless transmission of information by means of electromagnetic waves propagated through space. Importantly, the value of a radio increases tremendously when employed in a network to deliver services. We define a ‘wireless communication scene’ as the combination of the service, the propagation conditions, and the network situation. These three different elements of the wireless communication scene all feature a high degree of dynamics. 1. The service to be delivered. Advanced mobile terminals support next to the traditional ‘voice’, multi-media services ranging from data to video, and in the future maybe even 3D multimedia (for example for mobile gaming applications). The nature of the service highly influences the amount of information to be transmitted, and the constraints (e.g. on errors and latency) to be met. Also within one service, the dynamics can high. The actual rate of a video codec for example, very much depends on the specific images, and the correlation between frames. 2. The propagation conditions on the channel. The most prominent attribute is the path loss, which will define the average attenuation the signals undergo between transmitter and receiver. Next, the multi-path response encountered due to reflections of the waves, is an important characteristic. It will define the fading the signals encounter [11]. Handheld terminals can end up in different geographical environments. Moreover, as they can be used in mobile conditions, both the path loss and the fading characteristics can vary in time during communication. 3. The network situation. The ether is a shared medium, consequently a pool of multiple users can simultaneously send signals ‘in the air’. The multi-user traffic situation can be very dynamic as well, with the number of users and their communication intensity coming and going. The prominent dynamics in the wireless communication scene can and maybe should be adapted to, as will be explained in Section 2.4.

2 Statement made by IMEC’s T@MPO team working on a low power Turbo codec [16], turns out to be even more crucial for SDR systems and newer technology nodes.

24

2 Software Defined Radios

2.2.3 SDR Solutions: Scalability should be Everywhere When building SDR solutions, in first instance, functional flexibility is targeted. However, introducing reconfigurability in the different components of the radio typically comes at the risk of a power penalty. The key to enabling both low power and flexibility, is to target energy scalable SDRs. These radios feature a power consumption which scales down accordingly, if they do not need to deliver their maximum performance. Energy scalability in SDRs can and should be introduced everywhere in the system [18]: – – – –

Both in the analogue front-end and the digital baseband On different hierarchical levels in the design: platform, components, and circuits As well in the hardware as in the algorithmic solutions and the software Considered over different standards as well as within one standard

In the conception of green SDRs [17], functional flexibility and energy scalability are considered as firmly coupled design goals.

2.2.4 Exploit Dynamics and Scalability!

t

Communication conditions

Application requirements

A major challenge is to enable low energy SDR solutions, suited for handheld multimedia terminals and competitive with fixed hardware implementations. To make low energy terminals a reality, a two-step approach is advocated (Fig. 2.5) [17].

t

Cross-layer optimized Power/performance manager

Energy - scalable SDR Baseband Baseband

Front-end

Fig. 2.5 Reconfigurable and energy scalable radio solutions achieve low energy operation through cross-layer joint QoS and energy management

References

25

Traditional designs are still mostly tuned for the worst-case. By carefully scanning and following the exact (run-time) requirements without over-dimensioning the active part of the components, much energy can be saved. First, effective energy scalability is enabled in the design of the radio baseband and front-end. Secondly, the scalability is exploited to achieve low power operation by a cross-layer controller that follows at run-time the dynamics in the wireless communication scene. The approach illustrated above sketches the fundamental concepts of this book. Cognitive radios will extend the flexibility enabled by new reconfigurable radio architectures, by adding intelligent control solutions exploiting under-utilization of many licensed frequency bands. The design approach introduced above is clearly paving the way towards this challenging goal, as further explained in Chapter 7.

References 1. M. Dillinger, K. Madani, and N. Alonistioti, Software Defined Radio: Architectures, Systems and Functions, Wiley, Chichester, 2003. 2. J. Reed, Software Radio – A Modern Approach to Radio Engineering, Prentice-Hall, Upper Saddle River, NJ, 2002. 3. W. Eberle et al., A digital 80 Mb/s OFDM transceiver IC for wireless LAN in 5 GHz band, IEEE International Solid-State Circuits Conference, San Francisco, CA, Feb. 2000. 4. F.K. Jondral, Software-Defined Radio – Basics and Evolution to Cognitive Radio, EURASIP Journal on Wireless Communications and Networking, Vol. 2003, No. 3, pp. 275–283, 2005. 5. G. Desoli and E. Filippi, An Outlook on the Evolution of Mobile Terminals, CAS Magazine, second quarter 2006. 6. A. Sinha, A. Wang, and A.P. Chandrakasan, Energy Scalable System Design, IEEE Transactions on VLSI Systems, Vol. 10, No. 2, pp. 135–145, April 2002, Transaction on VLSI Systems, April 2002. 7. sca.jpeojtrs.mil. 8. www.sdrforum.org. 9. REM, Radio free Europe, Murmur, 1983. 10. Wireless World Research Forum (WWRF), Working Group 6 (Reconfigurability) http://wg6.ww-rf.org/. 11. T.S. Rappaport, Wireless Communications: Principles and Practice. Prentice Hall, Upper Saddle River, NJ, 1996. 12. www.IEEE802.org. 13. www.3gpp.com. 14. www.scc41.org. 15. J. Mitola et al., Cognitive Radio: Making Software Radios More Personal, IEEE Personal Communication, Vol. 6, No. 4, pp. 13–18, Aug. 1999. 16. B. Bougard et al., A Scalable 8.7-nJ/bit 75.6-Mb/s Parallel Concatenated Convolutional (turbo-) Codec, IEEE International Solid-State Circuits Conference, Feb. 2003, San Francisco, CA. 17. A. Dejonghe, B. Bougard, S. Pollin, J. Craninckx, A. Bourdoux, L. Van der Perre, and F. Catthoor, Green Reconfigurable Radio Systems: Creating and Managing Flexibility to Overcome Battery and Spectrum Scarcity, Signal Processing Magazine, May 2007. 18. L. Van der Perre et al., Architectures and Circuits for Software Defined Radios: Scaling and Scalability for Low Cost and Low Energy, ISSCC 2007, Feb. 2007, San Francisco, CA.

Chapter 3

Software-Defined Radio Front-Ends Scalable Waves in the Air

3.1 Introduction The ultimate dream of every Software-Defined Radio (SDR) front-end designer is to deliver an RF transceiver that can be reconfigured into every imaginable operating mode, in order to comply with the requirements of all existing and even upcoming communication standards. These include a large range of modes for cellular (2G–2.5G–3G and further), WLAN (802.11a/b/g/n), WPAN (Bluetooth, Zigbee, ...), broadcasting (DAB, DVB, DMB, ...) and positioning (GPS, Galileo) functionality. Obviously, all of them have different center frequency, channel bandwidth, noise levels, interference requirements, transmit spectral mask, etc. As a consequence, the performances of all building blocks in the transceiver must be reconfigurable over an extremely wide range, requiring ultimate creativity from the SDR designer. Reconfigurability is a requirement for SDR functionality, but often one forgets that it can also be an enabler for low power consumption. Indeed, once the flexibility is built into the transceiver, it can be used to adapt the performance of the radio to the actual circumstances, instead of those implied by the worst-case situation of the standard. Since linearity, filtering, noise, bandwidth, etc. can be traded for power consumption in the SDR, a smart controller is able to adapt the radio at runtime to the actual performance required, and hence can reduce the average power consumption of the SDR. In this chapter, several important innovations and concepts will be presented that bring this ultimate dream closer to reality. These include circuits for wideband LO synthesis, multifunctional receiver and transmitter blocks, novel ADC implementations, etc. The result of this all is integrated in the world’s first SDR transceiver covering the frequency range from 174 MHz to 6 GHz, implemented in a 1.2V 0.13 μ m CMOS technology.

Jan Craninckx IMEC, Leuven, Belgium, e-mail: [email protected] L. Van der Perre et al., Green Software Defined Radios, Series on Integrated Circuits and Systems, c Springer Science+Business Media B.V. 2009 

27

28

3 Software-Defined Radio Front-Ends

3.2 System-Level Considerations A first choice to be made is the radio architecture to be used. In the past decades, lots of studies and examples have been presented on heterodyne, homodyne, lowIF, wideband-IF, etc. architectures, all having certain benefits and problems for a certain application. Which one to choose? In view of SDR, this question becomes maybe a little bit easier to answer. Indeed, when the characteristics of all possible standards are taken into account, not a single intermediate frequency can be found that suits them all. And having multiple IFs increase the hardware cost of the SDR, which cannot be tolerated. So direct-conversion architectures are the right choice for the job. All of the well-known problems, such as DC offsets, I/Q mismatch, 1/f noise, PA pulling, etc. that have limited the proliferation of zero-IF CMOS radios into mainstream products have been solved in recent years, and it will enable the design of a low-cost front-end. A schematic vision of what the final SDR will look like is represented in Fig. 3.1. For low cost in a large-volume consumer market, the active transceiver core is implemented in a plain CMOS technology. It includes a fully reconfigurable direct conversion receiver, transmitter, and two synthesizers (for FDD operation). The functions that cannot be implemented in CMOS are included on the package substrate. These are primarily related to the interface between the active core and the antenna. They must provide high-Q bandpass filtering or even duplexing, impedance matching circuits, and power amplification. The remainder of this chapter will mainly focus on the transceiver implementation. The hard works starts with determining performance specifications of each block in the chain. The total budget for gain, noise, linearity, etc. must be divided over all

MCM substrate

CMOS IC

MEMS switches

Tunable matching

Frac-N PLL Frac-N PLL

DMQ VCO Distr. Tunable filtering

DMQ

Power amplifier

NoC controller

Fig. 3.1 Conceptual view of the SDR transceiver front-end

3.2 System-Level Considerations

29

blocks, ensuring that all possible test cases are covered, and this for every standard. Having very flexible building blocks of course helps a great deal, but making a smart system analysis at this point is crucial to obtain an optimal SDR solution. R tool called NETLISP has been developed to do this exerA custom Matlab cise, that ensures consistency during front-end design [14]. It takes in a netlist that describes all building blocks, with the performance characteristics and gain ranges, and simulates on a behavioral level the complete chain for a list of different test cases. All blocks can be modeled on varying levels of accuracy and complexity, but the integration of them in a single framework avoids inconsistencies between different levels of abstraction. Figure 3.2 shows a screenshot. The performance under all circumstances can thus be evaluated, and the building block performance can be tuned in order to fulfill all requirements. Gain ranges and signal filtering must be set such that the signal levels are an optimal trade-off between noise and distortion. Typically, four phases can be distinguished in the design flow [14]. 1. Exploration phase: Based on a generic behavioural model, plugged into endto-end system simulations, the effect of front-end non-idealities such as noise, nonlinearity, quadrature mismatch, offsets, etc. is analyzed. The resulting implementation loss curves allow the system designer to derive specifications for the different effect based on a budget for the complete link.

Fig. 3.2 System-level analysis tool

30

3 Software-Defined Radio Front-Ends

2. Cascade analysis phase: The cascade analysis distributes the specifications for the different non-idealities derived in the exploration phase over the different FE blocks. The building block models are equation-based, and allows easy evaluation of different architectures and Automatic Gain Control (AGC) or power control algorithms. 3. Design phase: This phase includes both the design of the analog circuits and the development of algorithms for compensation of the front-end non-idealities. Consistence of the simulated analog performance with the cascade analysis description is ensured by validation w.r.t. Verilog-A models that are automatically generated from the NETLISP description. 4. Verification phase: In this phase a verification of the designed FE is done. Complex and accurate models can be derived from the designed circuits, and they can be inserted in the cascade analysis to check the complete link performance. In a typical design scenario, there are some iterations between step 2 to 4. Although being a difficult excercise, the analysis can show that with the built-in flexibility also a software-defined radio can achieve stat-of-the-art performance very close to dedicated single-mode solutions. In the next sections we will go deeper into the design of some crucial building blocks.

3.3 Wideband LO Synthesis To generate all required LO signals in the range of 0.1–6 GHz, several frequency generation techniques have been proposed to relax the tuning range specifications of the voltage-controlled oscillator (VCO). They use division, mixing, multiplication or a combination of these [31]. However, to make these systems efficient in terms of phase noise and power consumption, the VCO tuning range still has to be maximized. The following section discusses the design of such a wideband VCO, whereas the architecture required to generate all LO signals will be discussed in Section 3.3.2.

3.3.1 3–5 GHz Voltage-Controlled Oscillator To reach the stringent phase noise specifications for todays mobile communication systems, most RF transceiver ICs use LC-VCOs. Frequency tuning of LC VCOs is often done by changing the capacitance value of the resonant tank using varactors. Switched or controlled inductor designs have been reported [28], but it remains difficult to cover the desired wide band continuously and to limit the deterioration of the phase noise performance caused by the insertion of these switches. Instead of using a single large varactor to tune the frequency, a mixed discrete/continuous tuning scheme is usually chosen [20]. A small varactor is used for fine continuous tuning whereas larger steps are realized by digitally switching

3.3 Wideband LO Synthesis

31

capacitors in and out of the resonant tank. This has two advantages: the VCO gain is lower, allowing easier phase-locked loop (PLL) design, and digitally switched varactors have a higher ratio between the capacitance in the on-state (CON ) and the capacitance in the off-state (COFF ). A higher CON /COFF ratio allows a larger VCO frequency tuning range. However, as the tuning range of a VCO is increased and exceeds the typical 20% range obtained in many designs, new problems and trade-offs appear that need a solution. In this design we have tackled the two main problems encountered in wideband LC-VCOs. First, the negative resistance required to maintain oscillation varies a lot over the frequency range, leading to significant overhead when a fixed active core is used. Secondly, the large variation of the VCO gain (KVCO ) across the whole tuning range creates problems for optimal and stable PLL design. Solutions are proposed for both problems.

3.3.1.1 Tank Loss Variations Required Negative Resistance In the target frequency range (