by Stéphane Téral, Executive Director, Research and Analysis, Mobile Infrastructure & Carrier Economics, IHS Markit
The 3GPP standards organization is moving quickly towards finalizing Phase 1 of 5G in Release 15. However, unlike in previous wireless generations that also had to meet new use case requirements, somewhat surprisingly there have been no fundamental changes made to the waveform. In part, this was done in order to meet the aggressive timeline, as many important 5G use cases and verticals were postponed to 5G Phase 2, with the study item efforts expected to begin in earnest in early 2018. Even more surprising is that most engineers in the wireless standards community assume that we can’t do much better than OFDM in approaching capacity. This is simply not the case.
Cohere Technologies has developed a breakthrough 2D modulation technique called OTFS (Orthogonal Time Frequency and Space) that mathematically unifies the classical 1D modulations of OFDM, TDMA and CDMA. The company has worked with many of the world’s leading operators (e.g. Cohere’s 3GPP RAN1 submissions were co-signed by AT&T, CMCC, DT, Telefonica, Telstra) and wireless academics (e.g., professors from Duke University, Stanford University, University of Southern California, University of Texas at Austin) to validate the technology and is proposing OTFS into 3GPP as an enhancement to OFDM for important new verticals. Cohere’s 3GPP RAN submissions can be found here: https://www.cohere-technologies.com/resources/
5G Phase 2 offers a real opportunity to examine critical new use cases, including next generation V2X, high-speed rail and MU-MIMO capacity-achieving performance improvements, which can only be gained through enhancements to the underlying OFDM waveform.
As a first example, new requirements for V2X include significantly enhanced reliability under high mobility scenarios. In particular, the vehicle-to-vehicle (V2V) requirements are unique in 3GPP: typically, vehicles will broadcast information from their sensors to other cars and thus cannot make use of the ubiquitous HARQ retransmission scheme currently used in cellular communications. Transmission must get through the first time, and with future autonomous driving applications this becomes an absolutely mission-critical requirement. It is imperative that the standards community explore all options to ensure the crucial V2V communications link is uncompromised.
A second important vertical not addressed in Release 15 is high speed rail, which requires massive amounts of bandwidth at speeds of up to 500 km/h – a regime where OFDM and LTE have demonstrably poor performance.
Finally, a key promise of 5G is the goal of approaching capacity in a MU-MIMO scheme. Optimal nonlinear precoding schemes to achieve this are not practical in OFDM, however they are ideally suited to OTFS.
The OTFS modulation technique carries the information QAM symbols over a new class of waveforms that correspond, via a transform called the Zak transform, to localized pulses in a two-dimensional signal representation called the delay-Doppler representation. The OTFS waveform is simultaneously localized in time like the TDMA pulse, localized in frequency like the OFDM tone and spread spectrum like the CDMA code. In this way, it combines the benefits of these three pillar waveforms of classical communication. In particular, it supports the optimal PAPR and resilience to phase noise of TDMA, simple channel model and flexible multi-user allocation of OFDM, and resilience to narrowband interference of CDMA, rendering it the absolute optimal air-interface for 5G. It has been shown that OTFS outperforms OFDM in all scenarios and independent academic research has confirmed the performance advantages. A lecture presentation on OTFS by Cohere’s CTO Dr. Ronny Hadani can be found here: https://www.cohere-technologies.com/resources/videos/
From a broader perspective, the OTFS waveform establishes a direct link between Radar and communication. It couples with the wireless channel in a way that directly captures the delay-Doppler radar image of the constituent reflectors, thus yielding a complete separation of the reflections according to their delay-Doppler characteristics and extracting full time-frequency diversity. This converts the dynamic wireless channel into a constant energy interaction, where all received QAM symbols have the same gain and all delay-Doppler diversity branches are coherently combined. As result, a wireless system designed over the OTFS waveforms has consistent performance which is independent of the channel conditions. In addition, the high-resolution delay-Doppler separation of the reflectors enables OTFS to approach channel capacity with optimal performance-complexity trade off through linear scaling of spectral efficiency with the MIMO order and robustness to Doppler and multipath channel conditions.
Given the level of spectral efficiency gains Cohere has demonstrated so far, I strongly believe OTFS is a serious 5G contender that will eventually displace OFDM.
OTFS is an enabler for realizing the full promise of MU-MIMO gains even in challenging 5G deployment settings where adaptation is unrealistic.
For more details on this groundbreaking fundamental waveform, including performance comparisons with LTE, see https://www.cohere-technologies.com/wp-content/uploads/2017/10/OTFS-Physics-White-Paper.pdf
Stéphane Téral is Executive Director, Research and Analysis, Mobile Infrastructure & Carrier Economics, IHS Markit.
With 26 years of experience in the telecommunications industry, Stéphane Téral is regarded as one of the top analysts in his field, having been the trusted advisor at some of the world's largest telecom providers and manufacturers. Stéphane now specializes in next-generation mobile infrastructure, voice over LTE and circuit-to-packet migration products and services, and adoption trends of service providers.
Dr. Ronny Hadani, CTO, Cohere Technologies and Dr. Anton Monk, VP Strategic Alliances and Standards, Cohere Technologies also contributed to this article.