Monday 11 January 2010

Technologies and Standards for TD-SCDMA Evolutions to IMT-Advanced

Picture Source: http://www.itu.int/dms_pub/itu-t/oth/21/05/T21050000010003PDFE.pdf

This is a summary of a paper from IEEE Communications Magazine, Dec 2009 issue titled "Technologies and Standards for TD-SCDMA Evolutions to IMT-Advanced" by Mugen Peng and Wenbo Wang of Beijing University of Posts and Telecommunications with my own comments and understanding.

As I have blogged about in the past that China Mobile has launched TD-SCDMA network in China and the main focus to to iron out the basic problems before moving onto the evolved TD-SCDMA network. Couple of device manufacturers have already started working on the TD-HSPA devices. Couple of months back, 3G Americas published a whitepaper giving overview and emphasising the advantages of TDD flavour of LTE as compared to FDD. The next milestone is the IMT-Advanced that is under discussion at the moment and China has already proposed TD-LTE-Advanced which would be compatible with the TD-SCDMA technology.

For anyone who does not know the difference between TDD, FDD and TD-SCDMA please see this blog.

The TD-SCDMA technology has been standardised quite a while back but the rollout has been slow. The commercial TD-SCDMA network was rolled out in 2009 and more and more device manufacturers are getting interested in the technology. This could be due to the fact that China Mobile has a customer base of over 500 million subscribers. As of July 2009 over 100 device manufacturers were working on TD-SCDMA technology.

The big problem with TD-SCDMA (as in the case of R99 3G) is that the practical data rate is 350kbps max. This can definitely not provide a broadband experience. To increase the data rates there are two different approaches. First is the Short Term Evolution (STE) and the other is Long Term Evolution (LTE).

The first phase of evolution as can be seen in the picture above is the TD-STE. This consists of single carrier and multi-carrier TD-HSDPA/TD-HSUPA (TD-HSPA), TD-MBMS and TD-HSPA+.

The LTE part is known as TD-LTE. There is a definite evolution path specified from TD-SCDMA to TD-LTE and hence TD-LTE is widely supported by the TD-SCDMA technology device manufacturers and operators. The target of TD-LTE is to enhance the capabilities of coverage, service provision, and mobility support of TD-SCDMA. To save investment and make full use of the network infrastructure available, the design of TD-LTE takes into account the features of TD-SCDMA, and keeps TD-LTE backward compatible with TD-SCDMA and TD-STE systems to ensure smooth migration.

The final phase of evolution is the 4G technology or IMT-Advanced and the TD-SCDMA candidate for TD-LTE+ is TD-LTE-Advanced. Some mature techniques related to the TD-SCDMA characteristics, such as beamforming (BF), dynamic channel allocation, and uplink synchronization, will be creatively incorporated in the TD-LTE+ system.

Some academic proposals were also made like the one available here on the future evolution of TD-SCDMA but they lacked the industry requirements and are just useful for theoretical research.

The standards of TD-SCDMA and its evolution systems are supervised by 3GPP in Europe and by CCSA (Chinese Cellular Standards Association) in China. In March 2001 3GPP fulfilled TD-SCDMA low chip rate (LCR) standardization in Release 4 (R4). The improved R4 and Release 5 (R5) specifications have added some promising functions including HSDPA, synchronization procedures, terminal location (angle of arrival [AOA]-aided location), and so on.

When the industry standardizations supervised by CCSA are focusing on the integration of R4 and R5, the N-frequency TD-SCDMA and the extension of HSDPA from single- to multicarrier are presented. Meanwhile, some networking techniques, such as N-frequency, polarized smart antenna, and a new networking configuration with baseband unit plus remote radio unit (BBU+RRU), are present in the commercial application of TD-SCDMA.

TD-SCDMA STE

For the first evolution phase of TD-SCDMA, three alternative solutions are considered. The first one is compatible with WCDMA STE, which is based on HSDPA/HSUPA technology. The second is to provide MBMS service via the compatible multicast broadcast single-frequency network (MBSFN) technique or the new union time-slot network (UTN) technique. The last is HSPA+ to achieve similar performance as LTE.

On a single carrier, TD-HSDPA can reach a peak rate of 2.8 Mb/s for each carrier when the
ratio of upstream and downstream time slots is 1:5. The theoretical peak transmission rate of a three-carrier HSDPA system with 16-quadrature amplitude modulation (QAM) is up to 8.4 Mb/s.

Single-carrier TD-HSUPA can achieve different throughput rates if the configurations and parameters are varied, including the number of occupied time slots, the modulation, and the transport block size in bytes. Considering the complexity of a terminal with several carriers in TD-HSUPA, multicarrier is configured in the Node B, while only one carrier is employed in the terminal.

In Rel-7 based TD-HSPA+, In order to match the performance of orthogonal frequency-division multiple access (OFDMA)-based TD-LTE systems, some advanced techniques are utilized, such as multiple-input multiple-output (MIMO), polarized BF, higher modulation and coding schemes (64-QAM is available), adaptive fast scheduling, multicarrier techniques, and so on. Theoretically, 64-QAM can improve performance by a factor of 1.5 compared to the current 16-QAM; for single-carrier the peak rate reaches 4.2 Mb/s, and three-carrier up to 12.6 Mb/s.

For the MIMO technique, double transmit antenna array (D-TxAA), based on the pre-coding method at the transmitter, has been employed in frequency-division duplex (FDD)-HSPA+ systems, while selective per antenna rate control (S-PARC), motivated by the Shannon capacity limit for an open loop MIMO link, has been applied in TD-HSPA+ systems.

TD-SCDMA LTE

The TD-SCDMA LTE program was kicked off in November 2004, and the LTE demand report was approved in June 2005. The LTE specified for TD_SCDMA evolution is named TD-LTE.

LTE systems are supposed to work in both FDD and TDD modes. LTE TDD and FDD modes have been greatly harmonized in the sense that both modes share the same underlying framework, including radio access schemes OFDMA in downlink and SC-FDMA in uplink, basic subframe formats, configuration protocols, and so on.

TD-LTE trials have already started last year with some positive results.

TD-SCDMA LTE+

IMT-Advanced can be regarded as a B3G/4G standard, and the current TD-SCDMA standard migrating to IMT-Advanced can be regarded as a thorough revolution. TD-LTE advanced (TD-LTE+) is a good match with the TD-SCDMA revolution to IMT-Advanced.

It is predicted that the future TD-SCDMA revolution technology will support data rates up to approximately 100 Mb/s for high mobility and up to approximately 1 Gb/s for low mobility such as nomadic/local wireless access.

Recently, some advanced techniques have been presented for TD-LTE+ in China, ranging from the system architecture to the radio processing techniques, such as multi-user (MU)-BF, wireless relaying, and carrier aggregation (CA).

For MU-BF see the paper proposed by Huawei, CHina Mobile and CATT here (http://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_55b/Docs/R1-090133.zip).

For Wireless Relaying see the ZTE paper here (http://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_56b/Docs/R1-091423.zip).

To achieve higher performance and target peak data rates, LTE+ systems should support bandwidth greater than 20 MHz (e.g., up to 100 MHz). Consequently, the requirements for TD-LTE+ include support for larger transmission bandwidths than in TD-LTE. Moreover, there should be backward compatibility so that a TD-LTE user can work in TD-LTE+ networks. CA is a concept that can provide bandwidth scalability while maintaining backward compatibility with TD-LTE through any of the constituent carriers, where multiple component carriers are aggregated to the desired TD-LTE+ system bandwidth. A TD-LTE R8 terminal can receive one of these component carriers, while an TD-LTE+ terminal can simultaneously access multiple component carriers. Compared to other approaches, CA does not require extensive changes to the TD-LTE physical layer structure and simplifies reuse of existing implementations. For more on Carrier Aggregation see CATT, LGE and Motorola paper here (http://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_56b/Docs/R1-091655.zip).

Finally, there are some interesting developments happening in the TD-SCDMA market with bigger players getting interested. Once a critical mass is reached in the number of subscribers as well as the manufacturers I wouldnt be surprised if this technology is exported beyond the Chinese borders. With clear and defined evolution path this could be a win-win situation for everyone.

3 comments:

Unknown said...

Very interesting insight. I wonder though how far operators will still go with the STE and whether leap frogging to TD-LTE has any drawbacks other than having to fall back on much lower speeds of 384kbps.

Are you aware of the Frequency reuse issues: will they be able to use N=1 while improving the latency and capacities!!!?
thanks again

Zahid Ghadialy said...

Hi Daniel,

It is not necessary to evolve to TD-LTE from TD-SCDMA only. You can evolve from FDD path as well. So as far as falling back is concerned, you fallback on the other technology present which can be HSPA+, HSPA or just HSDPA (FDD path).

Regarding the frequency reuse, I havent heard much discussions so I guess its not a problem but I am not qualified in that area to comment.

Mihai C. said...

hello. I want to find the standard for TD-SCDMA 3GPP. I like to know some technical aspects like emission power, sensibility for the base stations and also for mobile units, the antenna height, cable loss, the antenna gains, etc.
Can you help me?
thanks.