Home > News content

Restore the truth of the HUAWEI 5G standard

via:博客园     time:2018/5/11 19:13:27     readed:138

Overnight.

What is the contention

In May 11th, a number of similar problems suddenly appeared on the knowledge. What was the idea of how Lenovo did not give HUAWEI to the 5G standard had caused a chain reaction.

According to Lei Feng net, the incident is not a recent event, but in the two meeting of 3GPP organizations in October 2016 and November. The two meeting is about the voting of channel coding schemes.

Channel can be divided into data channel and control channel, control channel transmission control information, data channel transmission data, control channel and data channel can adopt different schemes respectively. Coding level, the code length can be divided into long code and short code, short code is more commonly used, long code frequency is relatively low, long code and short code can also use different coding schemes.

But in fact, there are only three coding schemes for the 5G standard channel scheme, namely LDPC, Polar and Turbo. It is worth noting that the three schemes were initially proposed by mathematicians instead of industrial chain manufacturers, and Qualcomm, HUAWEI and other manufacturers were doing industrial research along the way of mathematicians.

LDPC code was originally proposed by MIT professor Robert Gallager in 1962. It is the longest and most mature channel coding scheme, and LDPC code is applied in the WiFi standard.

Turbo code, proposed by French scientists C.Berrou and A.Glavieux, also has more than 20 years of application history, and Turbo codes are adopted in 3G and 4G standards, but 5G standards become the competitive battlefields of LDPC and Polar. According to the analysis, the main reason is that France is not the enemy of China and the United States in the contest of discourse power; from the technical point of view, the 5G standard requires the peak rate to reach 20Gbps. The relevant news points out that whether the Turbo code can meet such rate is still doubtful.

The Polar code was proposed by E. Arikan, a professor at the University of Turkey in 2007. E. Arikan is a student of Robert Gallager with LDPC code. The application time of Polar code is the shortest, but the technical index is relatively excellent.

In the end, the debate on the channel coding scheme becomes the competition between HUAWEI's Polar code and the high pass LDPC code, and there are some big powers in the game of discourse power.

Two ballots

The first meeting, which took place on October 15, 2016, was divided. Lei Feng learned from people familiar with the industry,

Lei Feng is further aware that the first version of the 5G international standard in June 2018 will be officially announced, but the uRRLC and mMTC standards are not formally formed for the eMBB scenes in the three 5G scenes.

According to the industry's current consensus, eMBB, uRRLC and mMTC are mainly applied to the following scenarios

  • EMBB: enhance mobile broadband, VR/AR, Ultra HD video and so on.

  • MMTC: large-scale machine class communication, for large-scale Internet of things and other businesses;

  • URLLC: ultra high reliability and low delay / delay, for driverless, industrial automation and other businesses.

In the first meeting, a more precise explanation is that the LDPC code, which is adopted by the high pass, is adopted as a long code encoding scheme for the data channel in the eMMB scenario. It does not contain all three scenarios, nor is it used for data channels and control channels. Similarly, only LDPC codes are used for long codes.

The person familiar with the matter also said that

The time was set back for the second meeting, held on November 17, 2016. Although the process was not clear, the results came as a further step forward in the LDPC code scheme. Long Code and short Code Encoding Scheme for data Channel in eMBB scenario Polar Code becomes a short Code coding Scheme for Control Channel in eMBB scene (Control Channel Code length is generally not more than several hundred, does not involve long codes, so there is a series of at that time

When asked if both long and short codes would reduce industry costs when using a Qualcomm package, the person said,

As shown in the two diagram below, Lenovo and Moto did not vote for Polar in the first meeting and voted for the second time.

data-pre-sourced=yesdata-pre-sourced=yes

data-pre-sourced=yes

Early this morning, Lenovo also issued a statement on the matter, saying

At the same time, Lenovo also said that it reserves relevant legal rights against malicious rumor, and expresses its attitude to support China's 5G.

After the first two parts of the narrative, the context of the event has become clear. As people familiar with the matter say,

The focus of most of the readers' anger is not that Lenovo voted for Qualcomm, but Lenovo is the only Chinese company in favor of manufacturers.

From an objective technical point of view, compared with the high pass LDPC coding scheme, the Polar encoding scheme of HUAWEI's main push has no technical advantage in the long code part, and the HUAWEI first also wants to get the short code encoding scheme.

At the same time, although the LDPC coding scheme is a high pass, in fact, HUAWEI and other manufacturers are involved, high Qualcomm investment is much earlier, and the technology accumulation is more. Polar code is a new coding scheme that has been developed in recent years, which has been favored by HUAWEI and other manufacturers. The advantage is HUAWEI and Qualcomm and so on. The final two schemes coexist, 3GPP is a compromise.

It was true, but after more than a year, Lenovo did not brush the ticket to HUAWEI again. Who could say it was just a natural fermentation?

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments