Non-Orthogonal Multiple Access (NOMA)

0 downloads 0 Views 2MB Size Report
Non-Orthogonal Multiple Access (NOMA):. Evolution towards 5G Cellular Networks. Zhiguo Ding. School of Computing and Communications. Lancaster ...

Non-Orthogonal Multiple Access (NOMA): Evolution towards 5G Cellular Networks Zhiguo Ding School of Computing and Communications Lancaster University A collaboration with Princeton, SWJTU, USTC, BUPT, FAU, AUTH, CMCC, Tsinghua, UT Dallas, XJTU, QMUL and Bell Labs

Outline:

• • • • •

NOMA basics MIMO-NOMA Cooperative NOMA Interplay between Cognitive Radio and NOMA Research challenges

NOMA: Basics Motivations for NOMA • Orthogonal multiple access has been used during the past – FDMA/TDMA/CDMA/OFDMA

• Dilemma to realize a better trade-off between system throughput and user fairness A promising solution is to break orthogonality • PD-NOMA, SCMA, PDMA, LPMA, and MUSA are based on NOMA What have happened so far • Included in various whitepapers for 5G (DOCOMO, METIS, NGMN, ZTE, SK Telecom, …) • Recently proposed to 3GPP-LTE (MUST)

NOMA: Basics Key ideas: • All the users are served at the same time, frequency and code • Users with better channel conditions get less power • Successive interference cancellation is used at the receivers

[1] Y. Saito, A. Benjebbour, Y. Kishiyama, and T. Nakamura, “System level performance evaluation of downlink non-orthogonal multiple access (NOMA),” in PIMRC 2013. [2] Z. Ding, Z. Yang, P. Fan and H. V. Poor, "On the Performance of Non-Orthogonal Multiple Access in 5G Systems with Randomly Deployed Users”, IEEE SPL, 2014.

NOMA: Basics Why NOMA is an ideal MA solution of 5G • Consider the following two scenarios – If one user only needs to be served with a low data rate, e.g. sensors. • The use of OMA gives the sensor more than it needs

– If one user has a very poor channel condition • The bandwidth allocated to this user via OMA is not used efficiently.

• The use of NOMA can accommodate the 5G requirements – High system throughput – Low latency – Massive connectivity

Standardization activities of NOMA - MUST: State of the art • At the 3GPP meeting in May 2015, it was decided to include MUST into LTE Advance • At the 3GPP meeting in August 2015, 15 forms of MUST have been proposed by – Huawei, Qualcomm, NTT DOCOMO, Nokia, Intel, LG Electronics, Sharp, Samsung, ZTE, Alcatel Lucent, …

• Finally accepted in December 2015 • For example, Huawei proposed three forms of NOMA – Non-Orthogonal Multiple Access (NOMA) – Semi-Orthogonal Multiple Access (SOMA) – Rate-adaptive constellation Expansion Multiple Access (REMA)

Standardization activities of NOMA - MUST: State of the art • All the forms of MUST can be categorized into three groups

MIMO-NOMA • Motivations – The advent of advanced cell phones, tablet computers and other hightech hand-held devices – MIMO offers excessive degrees of freedom to further improve the system throughput of NOMA

• Challenge – The key feature of NOMA is to exploit the difference between users’ channel conditions – In scenarios with single-antenna nodes, channels are scalar and it is easy to order the users based their channel conditions – In MIMO, channels are in form of matrices/vectors, which makes difficult to order users • It is not clear how to design optimal precoding/detection strategies.

MIMO-NOMA: Approach I • Assign different beams to different users

MIMO-NOMA: Approach I • The QoS is satisfied by forcing the beams to satisfy a predefined order – In the following example, the message to User 1, a user close to the cell edge, is to be decoded first, at all users – Then the message to User 2 can be decoded, after subtracting the message to User 1, – ...

[2] M. F. Hanif, Z. Ding, T. Ratnarajah and G. K. Karagiannidis “A Minorization-Maximization Method for Optimizing Sum Rate in Non-Orthogonal Multiple Access Systems, IEEE TSP, 2015.

MIMO-NOMA: Approach I • The QoS is satisfied by forcing the beams to satisfy a predefined order – In the following example, the message to User 1, a user close to the cell edge, is to be decoded first, at all users – Then the message to User 2 can be decoded, after subtracting the message to User 1, – ...

[2] M. F. Hanif, Z. Ding, T. Ratnarajah and G. K. Karagiannidis “A Minorization-Maximization Method for Optimizing Sum Rate in Non-Orthogonal Multiple Access Systems, IEEE TSP, 2015.

MIMO-NOMA: Approach I

[3] P. Xu, Z. Ding, X. Dai and H. V. Poor, “A New Evaluation Criterion for Non-Orthogonal Multiple Access in 5G Software Defined Networks”, IEEE Access, 2015

MIMO-NOMA: Approach I • The QoS is satisfied by forcing the beams to satisfy a predefined order – In the following example, the message to User 1, a user close to the cell edge, is to be decoded first, at all users – Then the message to User 2 can be decoded, after subtracting the message to User 1, – ...

[2] M. F. Hanif, Z. Ding, T. Ratnarajah and G. K. Karagiannidis “A Minorization-Maximization Method for Optimizing Sum Rate in Non-Orthogonal Multiple Access Systems, IEEE TSP, 2015.

MIMO-NOMA: Approach II • Decompose MIMO-NOMA to SISO-NOMA

MIMO-NOMA: Approach II - without CSIT (1/2) • Consider a BS with M antennas communicating with M groups of users, where there are K single-antenna users in each group • The BS sends

• Each user receives where P is an identity matrix. [4] Z. Ding, F. Adachi and H. V. Poor, “The Application of MIMO to Non-Orthogonal Multiple Access”, IEEE TWC, 2015

MIMO-NOMA: Approach II - without CSIT (2/2) • After applying a detection vector

where 𝒑1 =[1 0 … 0]T. • The detection vector is designed to satisfy

for different I and m. • As a result, MIMO-NOMA can be reduced to separated SISO-NOMA systems, where NOMA can be easily applied.

MIMO-NOMA: Approach II - with CIST (1/3) • We can also use the concept of signal alignment to decompose MIMO-NOMA into SISO-NOMA – Channel state information is needed at the BS.

• Consider a BS with M antennas communicating with M groups of users, where there are 2 single-antenna users in each group • The base station will send a precoded version of symbols

[5] Z. Ding, R. Schober, and H. V. Poor, “A General MIMO Framework for NOMA Downlink and Uplink Transmission Based on Signal Alignment”, IEEE TWC, 2016

MIMO-NOMA: Approach II - with CIST (2/3) • After applying the detection vector, each receiver observes

• Directly using zero forcing precoding, means

• To align the signals transmitted by the two users from the same pair • Now a zero forcing precoding vector only needs to satisfy

MIMO-NOMA: Approach II - with CIST (3/3) • Now the system models at the two users from the same pair can be changed as follows: and • Therefore an MIMO-NOMA system can be decomposed into these SISO-NOMA channels. • There are other ways to carry out decomposing, for example, to apply generalized eigenvalue decomposition, etc. • These decomposition methods can also be applied to – Uplink NOMA – SCMA and other 5G MA

MIMO-NOMA: Numerical results (1/2)

The targeted data rates for two users are 5 BPCU and 0.5 BPCU. U1 is in a disc with radius of 10m and U2 is in a ring with radius of 10m and 20m. M = N = 3. U2 gets ¾ of the power. The path loss exponent is α = 3. The noise power is −30dBm and the interference power is ρI = 0.

MIMO-NOMA: Numerical results (2/2)

The targeted data rates for two users are 5 BPCU and 0.5 BPCU. U1 is in a disc with radius of 10m and U2 is in a ring with radius of 10m and 20m. M = N = 3. U2 gets ¾ of the power. The path loss exponent is α = 3. The noise power is −30dBm and the interference power is ρI = 0.

Interplay between NOMA and Cognitive Radio • NOMA can be viewed as a special case of cognitive radio –The user with poorer channel state information (CSI) can be viewed as a primary user. –With orthogonal MA, bandwidth allocated to this user cannot be reused, which leads to poor spectral efficiency. –By using NOMA, a user with better CSI is admitted –Slight loss at the primary user, but significant improvement on system throughput

The base station is located at the origin, i.e., (0; 0). The user with the poor channel conditions, i.e., the primary user, is located at (5m; 0). The x-y plan denotes the location of the secondary user. The path loss exponent is 3. The transmit signal-to-noise ratio is 20dB. A fixed power allocation policy, (7/8, 1/8).

Interplay between NOMA and Cognitive Radio • The concept of cognitive radio is useful to NOMA –The advantage of NOMA can be clearly demonstrated by using cognitive radio –Cognitive radio can also be used to simplify the power allocation • NOMA power allocation needs to satisfy users’ quality of service • The use of cognitive radio provides an explicit expression for meeting such requirements.

The power allocation coefficient needs to satisfy a targeted data rate of 0.5 BPCU at the primary user. [6] Z. Ding, P. Fan and H. V. Poor, “On the impact of user pairing on NOMA”, IEEE TVT, 2015

• The application of NOMA is also beneficial to cognitive radio systems, – More secondary users can be admitted by using NOMA

Cooperative NOMA • Motivations – The feature of heterogeneous networks means that different users have different capabilities – There is redundant information inherited in NOMA systems • Users with better channel conditions know the information sent to the other users.

• Solution – Cooperative NOMA [7] Z. Ding, M. Peng and H. V. Poor, "Cooperative Non-Orthogonal Multiple Access in 5G Systems”, IEEE Communication Letters, 2015.

Cooperative NOMA + SWIPT • Motivation – To improve the reliability of the far NOMA users without draining the near users’ batteries

• consider the application of SWIPT to NOMA, where SWIPT is performed at the near NOMA users.

[8] Y. Liu, Z. Ding, M. Elkashlan, and H. V. Poor, “Cooperative Non-Orthogonal Multiple Access with Simultaneous Wireless Information and Power Transfer”, IEEE JSAC, 2016

Future directions • • • • • • • • • • •

Different variants of NOMA New coding and modulation for NOMA Hybrid multiple access User pairing/clustering MIMO and cooperative NOMA Interplay between NOMA and cognitive radio Imperfect CSI and limited channel feedback Security provisioning in NOMA Cross-layer optimization Implementation issues of NOMA Emerging applications of NOMA

Future directions • MIMO-NOMA – Can we combine massive MIMO with NOMA? • Solution: using spatial clustering • [9] Z. Ding and H. V. Poor, “Design of Massive-MIMONOMA with Limited Feedback”, IEEE SPL, 2016

– Can we do NOMA even if users have similar CSI? • Solution: using users’ dynamic QoS requirements • [10] Z. Ding, L. Dai and H. V. Poor, “MIMO-NOMA Design for Small Packet Transmission in the Internet of Things”, IEEE Access, 2016

Thank you for your attention. Questions? Z. Ding, Y. Liu, J. Choi, Q. Sun, M. Elkashlan, C-L. I and H. V. Poor, “Application of Non-orthogonal Multiple Access in LTE and 5G Networks”, submitted to IEEE Communication Magazine.

D. Fang, Z. Ding, Y. Huang, S. Shieh, G. Geraci, H. Claussen, and M. Zhang, Candidates for 5G Downlink Multiple Access, submitted to IEEE Communication Magazine

http://www.lancaster.ac.uk/staff/dingz/ Email: [email protected]

Suggest Documents