MIMO Enabled Efficient Mapping of Data in WiMAX Networks Phani Krishna P, Saravana Manickam R and Siva Ram Murthy C High Performance Computing & Networking Lab (HPCN) Department of Computer Science & Engineering Indian Institute of Technology Madras Chennai, India January 6, 2012 Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 1 / 15
Outline 1 Introduction to WiMAX Networks Multiple Antenna Techniques in WiMAX 2 Problem Statement Definition Solution Overview 3 Proposed Technique Transmitter Side Receiver Side Results 4 Conclusion Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 2 / 15
Introduction to WiMAX Networks Introduction to WiMAX Networks Uses Orthogonal Frequency Division Multiple Access (OFDMA) scalable from 1.25MHz to 20MHz Supports multiple antennas at both Base Station and Mobile Station Supports advanced antenna techniques such as spatial diversity, spatial multiplexing, beam forming, and collaborative spatial multiplexing Supports adaptive modulation and coding techniques Supports multiple QoS classes Popular wireless broadband solution Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 3 / 15
Introduction to WiMAX Networks Multiple Antenna Techniques in WiMAX Multiple Input Multiple Output (MIMO) Techniques Spatial Multiplexing Spatial Diversity Hybrid techniques Spatial Multiplexing s1 s2 s3 s4 Channel + AWGN Noise Antenna 1 Antenna 1 MS BS s 5 s6 s7 s8 Antenna 2 Antenna 2 Different symbols (data) are transmitted across different antennas in the same slot BS also utilizes multiple antennas to receive the transmitted symbols Improves the throughput of the network Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 4 / 15
Introduction to WiMAX Networks Multiple Antenna Techniques in WiMAX Spatial Diversity s1 s2 s3 s4 Channel + AWGN Noise Antenna 1 Antenna 1 MS BS s s s 3 s * * * * 1 4 2 Antenna 2 Antenna 2 Variant of the data transmitted at one antenna is transmitted at another antenna Increase in diversity order improves the reliability of transmitted data Hybrid Techniques Combination of diversity and multiplexing techniques can be used Trade-off between throughput and reliability is a well studied topic Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 5 / 15
Problem Statement Definition Problem Statement Propose a MIMO technique that provides reliability same as Diversity techniques, and throughput comparable to Multiplexing techniques, for uplink transmissions Probable Applications In high mobility conditions, Channel quality can not be determined perfectly all the time Currently, MIMO techniques are switched adaptively to maximize performance With the proposed solution, adaptive MIMO switching is not required Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 6 / 15
Problem Statement Solution Overview Solution Overview Transmit redundant data across different antenna only for first 2 slots Works only when data from one antenna is received with error, and the other stream is received without error Determine and Eliminate the error in the erroneous stream Result Higher throughput is achieved Proved that reliability of transmitted data is same as that of Diversity techniques Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 7 / 15
Proposed Technique Transmitter Side Proposed Technique s1 s 2 s3 s 4 Channel + AWGN Noise Antenna 1 Antenna 1 MS BS s 1 s 2 s5 s 6 Antenna 2 Antenna 2 Only the first two slots across both the antennas contain the same data Remaining slots at both the antenna will contain different data Both antenna transmit same data across all slots in Diversity techniques Both antenna transmit same data across few (2) slots in proposed technique Throughput is improved comparatively Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 8 / 15
Proposed Technique Receiver Side Challenges Improving reliability of the transmitted data across all slots Solution Works when one of the streams is erroneous and the other is received non-erroneous Determines the error in the erroneous stream by comparing with the correctly received stream, as same data is transmitted in first 2 slots of both streams Uses Kalman Filter at sub-carrier level to determine the error in erroneous stream The error determined remains constant for further slots, and data from the erroneous stream is recovered by removing the error Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 9 / 15
Proposed Technique Receiver Side Error Determination 1 1 2 3 4 Information Stream 1 2 m 48 Kalman Filter Estimates h 1 h 2 h 3 h 4 h 5 ofdm symbols sub carriers Information Stream 2 1 2 1 2 3 4 h 47 h 48 m 48 Estimate h 1 for sub-carrier 1 Update the estimate repeatedly for all the sub-carriers in one OFDM slot, for each sub-carrier separately The estimated error is constant for the given two streams, and Kalman filter updates converge to a value after an average of 40 estimations Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 10 / 15
Proposed Technique Receiver Side Assumptions Channel quality remains constant for each MS in a 5 ms frame Noise value remains in a Galois Field when a signal is processed Signal Level Block Diagram Antenna 1 RF SNR 1>T Yes SNR 1>T and SNR 2<T No SNR 1>T No Reject Signal No Yes Yes Proposed ADC Filter Demodulater Decoder No Yes Yes Antenna 2 RF SNR 2>T Yes SNR 2>T and SNR 1<T No SNR 2>T No Reject Signal Erroneous stream is convolved with inverse of the determined error signal, and the error is eliminated Traditional decoding techniques are used to retrieve data Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 11 / 15
Proposed Technique Results Network Performance Results 3.5e+06 1 3e+06 0.8 Throughput (Mbps) 2.5e+06 2e+06 Reliability 0.6 0.4 1.5e+06 Alamouti Proposed SingleAntenna 1e+06 10 15 20 25 30 35 40 45 50 Number of MSs 0.2 Alamouti Proposed SingleAntenna 0 10 15 20 25 30 35 40 45 50 Number of MSs Higher throughput is achieved Reliability equivalent to that of Diversity technique is achieved Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 12 / 15
Proposed Technique Results Kalman Filter Convergence 0.3 0.25 Percentage of Error 0.2 0.15 0.1 0.05 0 0 10 20 30 40 50 60 70 80 Number of Estimates The estimated error remains < 5% after 40 iterations Proved that Kalman filter is applicable Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 13 / 15
Conclusion Conclusion Throughput is maximized with highest possible reliability Works under the assumptions Channel quality remains constant in one OFDMA frame of 5 ms Data transmitted from only one antenna is received with out errors Future Work Relaxing the above assumptions to provide a solution Studying the characteristics of a mobile in high mobility situations Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 14 / 15
Conclusion Thank you! Phani Krishna P et al. (IIT Madras) Data Mapping in WiMAX Networks January 6, 2012 15 / 15