Full-duplex without Strings: Enabling Full- duplex with Half-duplex - PowerPoint PPT Presentation

Full-duplex without Strings: Enabling Full- duplex with Half-duplex Clients Karthikeyan Sundaresan, Mohammad Khojastepour, Eugene Chai, Sampath Rangarajan NEC Labs America MobiCom 2014

Full-duplex  Transmitting + receiving on same time- frequency resource FD Base Station  Key challenge: Self-interference SI  Several advancements in full-duplex design – Antenna + RF + digital cancelation – Three, two and single antenna designs FD Client – Co-existence with MIMO – Focus on peer-peer FD networks 1. “Achieving single channel, full duplex wireless communication”, J. Choi et. al. MobiCom, 2010. 2. “Experiment-driven characterization of full-duplex wireless systems”, M. Duarte et. al. IEEE Transactions on Wireless Communications, 2012. 3. “MIDU: Enabling MIMO Full-duplex”, E. Aryafar et al. MobiCom 2012. 4. “Full-duplex radios”, D. Bharadia et. al., Sigcomm 2013. 2

Distributed Full-duplex  Can we enable FD communication (2x multiplexing gain) with HD FD Base clients in a single cell? Station – Easier to embed FD functionality in BS/AP SI  Distributed FD UDI – Uplink from one client and downlink to another client HD Client HD Client  Key challenge: uplink-downlink interference (UDI) 3

Potential Solutions for UDI  Impact of UDI depends on topology d UDI (d – distance between BS and DL client)  Large impact for comparable distances 4

Potential Solutions for UDI  Impact of UDI depends on topology d UDI (d – distance between BS and DL client)  Implicit: leverage client separation Scaling to  Explicit: use side channels [Bai-Arxiv’12] MIMO?  Explicit: time-based interference alignment [Sahai-ITW’13]  Explicitly address UDI in the same channel in a scalable manner 5

Approach  Leverage spatial interference alignment to address UDI between HD clients – Use multiple antennas at HD clients y 2 y 1 y 1 ,y 2 – Pack interference in lesser x 1 ,x 2 x 2 dimensions x 3 x 1 x 4  Efficient: same channel y 3 ,y 4 x 3 ,x 4 y 3 y 4  Scalable: co-exist with MIMO x 4 x 2 x 3 x 1 DL UL  Deployable: only as challenging as MU-MIMO systems 6

Challenges  CSI overhead for UDI – More clients (dimensions), easier IA, but more overhead V 0 U 0 .... (N)  Constructing a feasible IA solution N streams N streams – MIMO precoders (V), receiver filers (U) U 1 at clients and AP .... .... V 1  Handling clients with heterogeneous antenna capabilities U 2 .... .... V 2 (N) (N)  Optimizing rate for the FD streams  FDoS: System that addresses above challenges to enable FD with HD clients 7

(1) Applying IA to FD Networks  Results – N even: 4 clients necessary to .... address UDI and enable 2N streams – N odd: 6 clients necessary N/2 N/2 (symmetric) U 1 – N odd: 5 clients necessary .... .... V 1 (asymmetric) N/2 N/2  Focus on symmetric FD networks U 2 .... .... V 2 – Constant overhead: CSI between 4 or 6 clients – Does not scale with N 1 1 .... .... U 3 V 3 8

(2) Constructing IA Solution  Receiver spatial dimensions – Desired (1:1) – Interference suppression (1:1) ..... – IA (1:many) 9

(2) Constructing IA Solution cyclic p+1 p+1 Construct a Determine IA ... ... feasible IAN solution p+1 p+1 ... ... q Select IA solution Find resulting IA for cyclic part solution for acyclic part ... ...  At most one cycle in IAN ... ... p+1 p+1  Closed-form IA solution acyclic p p cyclic ... ... acyclic p p K-q ... ...  2N streams achievable even with UDI for symmetric FD networks ... ... p p – With 4 (6) clients for N even (odd) ... ... 10

Example: N=5, 6 clients, 10 streams H 10 v 01 H 10 v 02 2 2 H 13 v 31 H 11 v 11 H 11 v 12 H 12 v 22 H 12 v 21 V 1 H 23 v 31 2 2 H 20 v 04 H 20 v 03 V 2 H 21 v 11 H 21 v 12 H 22 v 22 H 22 v 21 H 30 v 05 V 3 H 33 v 31 H 32 v 21 1 1 H 31 v 12 H 31 v 11 H 32 v 22 11

(3) Heterogeneous Clients  Clients with different number of antennas – Affects number of FD streams .... supported (N) ? streams ? streams  M+N streams achievable with FD .... ....  Different IA construction required (N) (M) – Combination of symmetric and asymmetric FD networks .... .... (N) (M) 12

Evaluation  Testbed – One AP and four clients (WARP nodes) with 2 or 4 antennas each – FD: SI cancelation based on prior works – Focus on UDI cancelation between UL and DL clients • Cancelation over 64 sub-carrier OFDM, 10 MHz channel – Experiments in indoor office environment  Baselines – HD system MU-MIMO (zero-forcing beamforming) – FD without UDI cancelation  Metric – SINR measurements, rate translation from SINR 13

Results (1) – UDI Suppression 10-15 dB 15-20 dB  10-20 dB of median UDI suppression out of 30 dB

Results (2) – Rate Performance  1.75-2x FD rate gain  1.5-2x gain over schemes not addressing UDI  Not addressing UDI can degrade performance to worse than HD

Conclusions  FD has potential to increase system capacity by 2x – All the more powerful if HD clients can be used  UDI is a key challenge in distributed FD networks  FDoS: a system that leverages spatial IA to address UDI – Theory and design of applying spatial IA for distributed FD – Incorporates practical considerations (overhead, rate, heterogeneity) – Demonstrates 1.5-2x gain in presence of UDI in practice  Next steps…  FD with HD clients in multi-cell networks

Thanks! 17

(3) Rate Optimization Jointly pick N/2 vectors UL each for V 1 ,V 2 that clients  Very challenging problem maximize rate of N UL (V 1 ,V 2 ) streams subject to IA – MU-MIMO precoding on DL and UL coupled through IA between DL-UL Fix receive filter U 0 AP  Modular design for AP from UL (U 0 ) optimization – De-couple IA from MU-MIMO precoder (rate) optimization – Retain structure of IA solution for UDI – Optimize DL and UL MU-MIMO precoders Given V 1 ,V 2 , pick receiver DL filters U 1 ,U 2 orthogonal to given IA solution clients sub-space spanned by (U 1 ,U 2 ) interference  Distributed realization of IA solution – Overhead reduced further by half Pick precoder V 0 at AP AP to maximize rate (V 0 ) of N DL streams 18

FDoS Operations Client and mode (FD Distributed computation Estimate CSI for UL, vs. HD) selection based of solution (AP DL and UDI channels on multiplexing gain, broadcasts only one with reduced feedback scheduling policy precoder) AP solicits/delivers block AP coordinates joint UL ACKs similar to and DL transmissions MU-MIMO during FD 19

Results (3) – Heterogeneity .... (4) 2 streams 4 streams .... .... (4) (2) .... .... (4) (2)  6 streams sent in heterogeneous set-up  Leverages heterogeneous antenna capabilities effectively

Results (4) - Scalability (a) With rate optimization (b) Without rate optimization  Evaluated FDoS design for larger (even/odd) N  FD gains scale and more pronounced with rate optimization 21