LTE eNodeB Prototype Based on GPP
Platform
Niu Kai
#1
, Sun Jianxing
#2
, He Zhiqiang
#3
, Kok Keong Chai
*4
#
Key Lab of Universal Wireless Communications,
Ministry of Education Beijing University of Posts and Telecommunications
Beijing 100876, China
1
niukai@bupt.edu.cn,
2
sunjx1017@126.com,
3
hezq@bupt.edu.cn
*
School of Electronic Engineering and Computer Science
Queen Mary University of London, UK
4
Michael.Chai@eecs.qmul.ac.uk
Abstract— This paper presents a real-time LTE signal
processing system based on general purpose processor
(GPP) platform. LTE is one of the wideband wireless
communication systems with strict requirements of
performance and real-time. To meet these requirements,
high-efficiency implementation methods such as SIMD
instructions and multicore processing are applied to the
system for accelerating the processing. Finally, we
achieved an effective LTE eNodeB prototype on
commodity PC platform with a low processing latency.
The prototype supports most functions of MAC and
PHY and reveals the feasibility of realizing real-time
high-throughput signal processing systems based on
GPP platform.
Keywords— LTE, signal processing, general purpose
processor, optimization, real-time.
I. INTRODUCTION
SDR (Software defined radio) [1] has taken quite
great development in the past years. Compared to the
traditional dedicated hardware platforms, this
technology tries to develop a fully programmable
wireless communication system on general purpose
computing platforms. The fast growing ability of
general purpose processors (GPPs) and digital signal
processors has made the implementation of SDR
platforms feasible to meet the processing
requirements of modern wireless communication
protocols. An early example of SDR platform is
SoftMAC [2] implement customized MAC protocols,
but PHY remained unchangeable. GNU radio
platforms [3] such as the Universal Software Radio
Peripheral (USRP) based on field programmable gate
arrays (FPGAs) can only sustain low-speed wireless
communication due to both the hardware and
software constraints [4]. WARP [5] and KUAR [6]
started the software based implementation of some
communication systems with the help of PowerCPU
cores. However, their platforms were based on
FPGAs and Digital Signal Processors (DSP) with
high hardware cost, difficult developing languages,
poor software development environment of support
and limited debugging tools.
Sora [7] based on commodity GPP architecture has
enabled users to develop medium-speed wireless
implementations, such as the IEEE 802.11a/b/g PHY
and MAC. This GPP-based SDR platform has
advantages: easy to develop and transplant, good
flexibility and simple to switch among the existing
air-interface standards. However, for LTE
communication system it is still
tremendous challenge to meet the real-time
requirement of the high speed digital baseband signal
processing on GPPs.
As compared to the former communication
systems on GPP platform (Sorawifi), the main
challenge for implementing SDR TD-LTE system is
the complexity of large amount of computation task.
As the key techniques of LTE, Orthogonal Frequency
Division Multiplexing (OFDM) and Multiple-Input
Multiple-Output (MIMO) improve the performance
of 4G mobile system whereas these techniques
significantly increase the complexity and amount of
the computation task. Traditional implementation
methods are not sufficient for the system to meet the
real-time requirement of the protocols at this point.
Therefore, there is a need of new optimization
methods for signal processing realization to make the
real-time system feasible.
This paper will exhibit a prototype and introduce
the optimization methods for 2X2 MIMO-LTE
implementation. The prototype consists of LTE
eNodeB signal processing system supporting most of
MAC and PHY functions such as HARQ, adaptive
modulation and coding (AMC), MIMO detection,
channel estimation and channel coding. The readers
can refer to 3GPP protocols [8-11] for the details.
Physical Uplink Shared Channel (PUSCH), Physical
Uplink Control Channel (PUCCH) for uplink and
Physical Downlink Shared Channel (PDSCH),
Physical Broadcast Channel (PBCH), Physical
Control Format Indicator Channel (PCFICH),
Physical Downlink Control Channel (PDCCH),
Physical Hybrid-ARQ Indicator Channel (PHICH)
for downlink are available in our system.
GC'12 Workshop: International Workshop on Cloud Base-Station and Large-Scale Cooperative Communications
978-1-4673-4941-3/12/$31.00 ©2012 IEEE 279