Moore’s Law and the Path to LTE
By Jan Whitacre, Agilent Technologies, Inc.

This article will explain how increasing processing power has changed the way successive generations of cellular communications systems treat the transmissions between base station and subscriber. Moore’s Law – transistor count on integrated circuits will double every two years for the foreseeable future – and Intel executive David House’s prediction that system performance would continue to double every 18 months have radically affected the capability expected of successive generations of cellular system, and have revolutionized the way we communicate today.

Less than 25 years ago, at the beginning of the digital cellular age, GSM – the now ubiquitous cellular standard – included elements that minimized the effects of multipath transmission. These mainly focused on the physical design of the RF network. The combination of frame length, reduced handset power compared to earlier analog systems (ie reduced cell sizes) and support for antenna diversity techniques at both transmitter and receiver made for greater robustness of connection in a world that focused on the integrity of circuit-switched voice calls. However, multipath effects, particularly in built-up areas, caused dropped calls and customer dissatisfaction. The “cure” was higher cell density and more handovers between cells, resulting in a more complex network infrastructure and a correspondingly higher overhead. Agilent’s phase and frequency error measurements gave designers a way to look at the complex dynamics of bursted GMSK signals, and thus minimize interference between users.
Later versions of the GSM specification included support for packet data and “always on” connections, and higher integration, smaller size and improved battery life made data-capable handsets possible, but there was little interest from subscribers for the services proposed by the network operators, in part due to the low data rates that were possible.

Third generation networks, based on CDMA technology, were designed from the beginning to provide viable data services. Increased processing power made real-time generation and reception of signals in the same frequency and time space, based on unique digital codes, fact rather than theory. Wider transmission bandwidth, along with the diversity techniques used in GSM, means they are less susceptible to fading effects. In addition, these systems use rake receivers, where a number of sub-receivers decode the energy received over different paths and combine it to optimize signal reception. Agilent contributed Code Domain power measurements – a display of power in each of the code channels – and system simulation of elements like rake receivers to give designers new insights into system behavior. For the first time, we see multipath transmission treated as a benefit rather than a curse. In some of the latest High Speed Packet Access (HSPA) implementations, mobile network operators are offering headline data rates up to 14Mbps.

New for 2010/2011 is 3GPP LTE, the Long Term Evolution of the cellular network. First specified in Release 8 of the Third Generation Partnership Project (3GPP) standards, it places huge demands on processing power. It has 6 optional RF bandwidths and uses a variant of Orthogonal Frequency Division Multiplexing (OFDM), an FFT-based multi-carrier RF architecture. (As an aside, OFDM was first proposed as a theory in the late 1950s, but had to wait almost 50 years for processing power to make it real in wireless systems.) Its goals, compared to 3G HSPA systems are:

1. 2 - 4 x spectral efficiency
2. 10 x improvement in latency (turnaround time)
3. headline data rates up to hundreds of megabits per second using advanced multi-antenna techniques
4. a data-centric, packet-only, system with voice traffic carried using VoIP.
5. support for both Frequency Division (FDD) and Time Division Duplex (TDD) operation

LTE is being developed from the start with flexible deployment in mind, and its FDD and TDD modes of operation will be much better integrated than was the case with earlier standards. A factor of four data rate increase is provided by the 20 MHz system bandwidth and the other increases come from the use of multiple antenna technology, which is also being designed in from the start. These high performance targets have limits, however. The peak rates will only be seen at pedestrian speeds, with system performance dropping off as speeds increase. Packet-only operation offers lower network costs, but creates some challenges for integrating LTE voice with legacy circuit-switched systems. The target for latency has brought significant change to the network topology. Higher processing power in the base station allows significantly more intelligence to be placed nearer to the subscriber rather than further back into the network – for example, data acknowledgement and channel reconfiguration.

Multiple antennas can be used in a variety of ways. In addition to normal transmit and receive diversity, LTE uses Spatial Multiplexing (or Multiple Input, Multiple Output (MIMO)) techniques where each transmit antenna sends different data to the subscriber, and each receiver will see a mixture of the signal from each of the transmitters as seen in Figure 2. The receivers use complex matrix mathematics to demodulate the signals and recover the original data.

The subscriber device measures reference signals in the transmitted data to estimate the real-time characteristics of the “channel” between transmitters and receivers. A passing bus or walking behind a building will have significant effects over both long and short term channel conditions. In closed loop systems, the subscriber device must measure the Received Signal Strength of the base station reference symbols, attempt to decode the data, and provide positive or negative acknowledgements for each data block. Based on these, it must also derive a Channel Quality Indicator (CQI) value, which the base station will use to determine throughput to the device by setting the modulation scheme and other transmit parameters. For cells with more than one antenna, it must also provide Rank Index, indicating which type of multi-antenna technique the device believes to be most suitable for the current conditions.

MIMO and OFDM are a good pair, as the digital processing required is more straight forward than CDMA systems. The new capability required by developers is the ability to stress-test these multiple-receiver systems and ensure they can decode and recombine the data from their separate inputs. Agilent’s PXB receiver tester provides the multiple signals and real-time fading required to fully evaluate the receiver systems in LTE subscriber devices, be they mobile phones or USB computer accessories.

Agilent Technologies has been involved in the standards-setting process of cellular technologies from the early analog predecessors of GSM to the latest preliminary discussions on LTE Advanced – the planned second phase of LTE – and has contributed vital knowledge to ensure specified parameters can be measured. As a result, and of course with our own help from Moore’s Law, the company has been able to provide the advances in measurement science needed for simulation, emulation and measurement during early product development, and the tools manufacturers and operators need to ensure that devices from different manufacturers meet their specifications and will interoperate successfully.

About the Author
Janine R. Whitacre is an Application Engineer for LTE with Agilent’s Electronic Measurement Group. Janine received her BS degree in electrical engineering from the University of Wisconsin, Madison, in 1979. She has over 10 years of experience working on emerging cellular technologies and is currently focused on application engineering for LTE.

Agilent Technologies, Inc.
www.agilent.com
Email this article to a friend!