Large Intelligent Surfaces - Massive MIMO Evolution or Revolution? · 2019. 12. 5. · Large Intelligent Surfaces - Massive MIMO Evolution or Revolution? Rui Dinis12 1Instituto de

Large Intelligent Surfaces - Massive MIMO Evolutionor Revolution?

Rui Dinis12

1Instituto de Telecomunicações

2Nova University of Lisbon

Outline

1 Motivation

2 MIMO

3 Massive MIMO

4 LISConceptChallenges

5 Conclusions

Lisbon, Nov. 25, 2019 2/15

Digital Communications

• Low error rates• Higher and higher bit rates• Spectral efficiency [bps/Hz]

• Power savings

Lisbon, Nov. 25, 2019 3/15

Channel Capacity

C = B log2(1 + SNR)

• Channel coding• Turbo codes• LDPC codes• Polar codes

• Equalization• MLSE• OFDM• SC-FDE

• Synchronization and channel estimation

Lisbon, Nov. 25, 2019 4/15

MIMO Channel

C = log2(

det(

I + SNRNT x

HHH))

• Channel capacity grows with the number of antennas• Gain relatively the SISO case upperbounded by min(NT x, NRx)• Suitable for OFDM and SC-FDE schemes• Optimum receiver too complex• Practical receivers based on MMSE with excellent

performance/complexity trade-offs

Lisbon, Nov. 25, 2019 5/15

Massive MIMO

• Conventional MIMO schemes suitable for systems up to about 8 × 8• Massive MIMO not a scaled version of MIMO!• Low complexity implementations (low resolution DACs and ADCs,

strongly nonlinear amplifiers, simplified equalization/pre-coding, etc.)• Common elements (RF chains, mixers, DAC/ADC, etc.)• Channel estimation challenges (e.g., pilot contamination)

Lisbon, Nov. 25, 2019 6/15

LIS - Large Intelligent Surfaces

• Evolution of massive MIMO• Much more antenna elements• Short range• Near field communication• LoS communication

Lisbon, Nov. 25, 2019 7/15

LIS Access

• Antennas switched on and off according to user position and/or userrequirements

• Resource allocation at the space domain

Lisbon, Nov. 25, 2019 8/15

LIS for Positioning

• Antennas with different RSS and/or AoA/AoD• Accurate positioning

Lisbon, Nov. 25, 2019 9/15

LIS for Communication

• Communication aided by positioning information• Low complexity transmission and detection schemes• Huge capacity and coverage gains• Robustness to interference and imperfections

Lisbon, Nov. 25, 2019 10/15

LIS for Energy Harvesting

• Beamforming to compensate losses in energy harvesting• Better range and/or energy harvesting efficiency than traditional

techniques• Ranges of 1m or more

Lisbon, Nov. 25, 2019 11/15

Transceiver Design

• Need for very low complexity transceivers• On/off approaches• Beamforming• Skip equalizers?• Interference cancellation• Low resolution DAC/ADC (1 bit quantizers?)• Low complexity amplifiers (saturated or even switched amplifiers)

Lisbon, Nov. 25, 2019 12/15

Channel Estimation

• Too many channel to estimate• Parameterized channel models• Position-aided channel estimation• Channel tracking

Lisbon, Nov. 25, 2019 13/15

Resource Allocation

• Space-domain resource allocation• Aided by position information• LIS split in panels

• Many antennas per panel• Small number of outputs per panel• A user can be associated to several panels

Lisbon, Nov. 25, 2019 14/15

Conclusions

• Path from SISO to LIS• LIS with very high potential• Challenges• It is possible!

Lisbon, Nov. 25, 2019 15/15

MotivationMIMOMassive MIMOLISConceptChallenges

Conclusions