Downlink Multi-User MIMO for IEEE m

Transcription

1 Downlink Multi-User MIMO for IEEE 80216m Sivakishore Reddy Naga Sekhar Centre of Excellence in Wireless Technology 2013

2 Outline 1 Introduction 2 Closed Loop MU-MIMO 3 Results 4 Open Loop MU-MIMO 5 Results 6 Conclusions Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

3 Outline 1 Introduction 2 Closed Loop MU-MIMO 3 Results 4 Open Loop MU-MIMO 5 Results 6 Conclusions Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

4 MIMO Introduction Figure: MIMO Systems Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

5 SU-MIMO vs MU-MIMO Single User MIMO : Promises reliability and channel capacity through diversity gain and rate maximization Mainly acts as physical(phy) layer performance booster Multi User MIMO : Uses spatial degrees of freedom to schedule multiple users to simultaneously share the spatial channel Optimum design strategy to increase the system capacity Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

6 Requirements of MU-MIMO : Requires Channel State Information Partial CSI Codebook based precoding Requires more complex scheduling algorithms Requires advanced transceiver methodologies Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

7 Outline 1 Introduction 2 Closed Loop MU-MIMO 3 Results 4 Open Loop MU-MIMO 5 Results 6 Conclusions Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

8 System model User1 data User 2 data User N udata PF SCHEDULER x 1 x 2 x K PRECODER (W) 1 2 Nt 1 Nr 1 Nr User 1 User 2 ^ x 1 ^ x 2 1 Nr User K ^ x K Feedback: CQI, PMI Figure: MU-MIMO System model Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

9 Notation N t : Noof antennas at the BS N r : Noof antennas at the MS/user N u : Noof users contending for resource unit K : Noof users served per resource unit H : N r N t desired channel matrix G : N r N t interfering co-channel matrix W : N t K precoder v : N t 1 precoding vector chosen from codebook I : Number of strong interferers Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

10 About Codebook based precoding In 16m Codebook is defined for a given MIMO setup W = [ v 1 v 2 v K ] N t K The N r 1 received signal vector at the k th MS : y k = H k Wx+n k, P = K H kv k x k + k {1,2,,K} K H k v i x i +n k P K i k,i=1 Desired signal IUI Noise x = [ P K x 1 P K x 2 P K x K ] T K 1 Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

11 CL-MUMIMO PROBLEM STATEMENT How should each user make his choice for the PMI? How should each user model his CQI? How should the BS use this feedback-info to schedule multiple users per data region? Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

12 Choose PMI to minimize MSE MUMIMO Equation : y k = H k Wx+ MSE for i th precoding vector is : MSE i = E(x s ˆx s ) 2 I G i W i z i +n k i=1 If W is unitary, = σx 2 s v ih [ k Hk WW H k +I cov +σ 2 I ] 1 Hk v i I I cov = G i W iw i G i i=1 MSE i = σx 2 s v [ i H k Hk H k +I cov +σ 2 I ] 1 Hk v i ( ) PMI = arg min MSE i i subcarriers Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

13 CQI modeling Receiver uses an MMSE filter b, b k y k = b k H kv k x k + CQI = K i k,i=1 b k H kv i x i + I b k G iw i z i +b k n k i=1 b k H kv k 2 K i k,i=1 b kh k v k 2 +b k(i cov +σ 2 I)b k where I cov = I i=1 G iw iw i G i = I i=1 G ig i Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

14 MUMIMO PF-Scheduler Find all possible pairs of users who have reported orthogonal PMI Find sum-pf-metric for each pair Schedule pair with maximum sum-pf-metric Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

15 Outline 1 Introduction 2 Closed Loop MU-MIMO 3 Results 4 Open Loop MU-MIMO 5 Results 6 Conclusions Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

16 Unitary vs Non-Unitary Precoders The covariance term which influence CQI : I cov = I i=1 G i W i W i G i If the precoder W is not unitary, the interference level may change from one frame to another frame as the precoders used changes So the CQI of 2x2 is stable compared to 4x2 as the precoders of 4x2 system are non-unitary Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

17 CQI Stability 1 08 Comparision of Histograms 2x2 4x2 Histogram CQI(n 1) PPSINR(n) in db Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

18 2x2 vs 4x2 1 CDF of User Throughputs 08 2x2 Ideal I cov 06 4x2 Ideal I cov CDF User Throughput (in kbps) Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

19 Practical Implications in Closed loop system The mimo modes of interfering users can change in closed loop region The calculation of CQI which involves I cov cannot be perfectly determined The PMI decision term also involves I cov is also not perfect In demodulation, if the mimo mode of CCI is different it is difficult to even estimate the interference covariance at the reciever Thus degrading the performance further The PMI and CQI fedback are IMPERFECT and reduces the capacity Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

20 CL-MUMIMO 1 CDF of User Throughputs CDF Practical 4x2 Practical 2x2 2x2 Ideal I cov 2x2 Scaled I cov 4x2 Ideal I cov 4x2 Scaled I cov User Throughput (in kbps) Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

21 CL-MU-MIMO vs OL-SU-MIMO 3 Sector Spectral Efficiency in bps/hz OL SU MIMO,2x2 CL MU MIMO,2x #Active users per sector Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

22 Outline 1 Introduction 2 Closed Loop MU-MIMO 3 Results 4 Open Loop MU-MIMO 5 Results 6 Conclusions Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

23 Open Loop MU-MIMO System Model User1 data User 2 data User N udata PF SCHEDULER x 1 x 2 x K PRECODER (W) 1 2 Nt 1 Nr 1 Nr User 1 User 2 ^ x 1 ^ x 2 1 Nr User K ^ x K Feedback: CQI, PSI Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

24 Open Loop Multi-User MIMO Precoders are fixed a priori W = [ v 1 v 2 v K ] N t K Each user feedbacks preferred stream index (P SI) and CQI PSI {1,2,,K} PF scheduler will serve set of K users who feedback PSIs {1,2,,K} Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

25 Open Loop Region Inside Open Loop Region : All the base stations will use same MIMO mode Creates stable interference environment Estimation of CCI is easy and accurate Covariance matrix is calculated using estimated precoded channels Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

26 Received signal The N r 1 received signal vector at the k th MS : y k = H k Wx+ = 8 G ik Wx ik +n k, i=1 P K H kv k x k + P K K i k,i=1 H k v i x i + k {1,2,,K} 8 G ik Wx ik +n k Desired signal IUI CCI Noise i=1 x = [ P K x 1 P K x 2 P K x K ] T K 1 Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

27 Covariance matrix and MMSE filter The Covariance matrix of CCI (K CCI ) : K CCI = ( ) 8 P G ik WW G ik K i=1 1 N r MMSE filter equation for k th user : [ b k = (H k v k ) H k H k + K P K CCI + KN ] 1 o P I N r where H k = H k W Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

28 SINR Calculation ˆx k = b k y k = P K b kh k v k x k + P K K 8 b k H k v i x i + b k G ik Wx ik +b kn k i k,i=1 i=1 ( P ) K bk H k v k 2 SINR k = IUI k +CCI k + b k 2 N o ( ) P K IUI k = b k H k v i 2 K i k,i=1 CCI k = b k K CCI b k,l Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

29 CQI and PSI Computation W = [ v 1 v 2 v K ] N t K SINR k,l = ( P K ) bk H k v l 2 IUI l +CCI k + b k 2 N o PSI k = arg max l SINR k,l CQI k = SINR k,psik Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

30 Open Loop Single-User MIMO Single user with single stream is scheduled No orthogonal pairing problem Better Cell edge user performance than MU-MIMO Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

31 Outline 1 Introduction 2 Closed Loop MU-MIMO 3 Results 4 Open Loop MU-MIMO 5 Results 6 Conclusions Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

32 OL-MU-MIMO vs OL-SU-MIMO Sector Spectral Efficiency in bps/hz OL MU MIMO,2x2,K=2 OL SU MIMO,2x #Active users per sector Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

33 OL-MU-MIMO vs OL-SU-MIMO contd 025 Cell Edge user spectral efficiency in bps/hz OL SU MIMO,2x2 OL MU MIMO,2X2,K=2 CL MU MIMO,2X2,K= #Active users per sector Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

34 OL-MU-MIMO, CL-MU-MIMO and OL-SU-MIMO Comparison CDF OL MUMIMO,actual covar CL MUMIMO,scaled covar OL SU MIMO sector SE,cell edge SE CL: 23655,0071 OL: 22754,0062 SU: 23857, frames,tc=3000, 2X2,2 users Doppler=7 Hz Throughput in Kbps Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

35 Sector Spectral Efficiency variation for OL-MU-MIMO : N u 4 4,K = 2 4 4,K = 3 4 4,K = Table: Sector Spectral Efficiency for 4 4 OL-MU-MIMO Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

36 Convergence of OL-MU-MIMO and CL-MU-MIMO When there are large number of active users per sector both OL-MU-MIMO and CL-MU-MIMO will perform similarly Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

37 Outline 1 Introduction 2 Closed Loop MU-MIMO 3 Results 4 Open Loop MU-MIMO 5 Results 6 Conclusions Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

38 Conclusions MU-MIMO system achieves higher throughput when compare to single user system MU-MIMO system has slightly more FER when compare to single user system Optimum number of users (with single stream per user) that can be scheduled in a dataregion to achieve maximum throughput is equal to min(n t,n r ) FER and throughput of the system improves as min(n t,n r ) increases Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

39 THANK YOU Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40

40 ANY QUESTIONS? Sivakishore Reddy Naga Sekhar (CEWiT) MU-MIMO for IEEE 80216m / 40