MIMO Systems and Energy Efficiency of Wireless Sensor Networks

Transcription

1 MIMO Systems and Energy Efficiency of Wireless Sensor Networks Olivier BERDER ENSSAT, Université de Rennes1 INRIA/IRISA EPC CAIRN December 10, 2012 HDR Defense, Olivier Berder, 10/12/12 1 / 42

2 HDR Defense, Olivier Berder, 10/12/12 2 / 42 Wireless Sensor Networks Wide range of Wireless Sensor Network (WSN) applications Health, buildings and agriculture monitoring, defense, etc 2Be per year market until 2022 Set of smart radio nodes generating and relaying messages Ad Hoc, fault tolerant networks Low cost, low traffic and low power

3 Which parts of a WSN node are energy consuming? A WSN node is a typical embedded system Classical power budget Radio: mw Processor: 5-10 mw HDR Defense, Olivier Berder, 10/12/12 3 / 42

4 How to design an energy efficient WSN platform? 1. Decrease Transmit Power Efficient signal processing Error detection and correction 2. Optimize radio activity MAC protocols 3. Optimize hardware architecture Co-processing, DVFS, power-gating Energy harvesting Goal of future WSNs: reach energy autonomy! HDR Defense, Olivier Berder, 10/12/12 4 / 42

5 HDR Defense, Olivier Berder, 10/12/12 5 / 42 Energy Optimization Methodology

6 HDR Defense, Olivier Berder, 10/12/12 5 / 42 Energy Optimization Methodology

7 HDR Defense, Olivier Berder, 10/12/12 5 / 42 Energy Optimization Methodology

8 HDR Defense, Olivier Berder, 10/12/12 5 / 42 Energy Optimization Methodology

9 Research themes at a glance Energy modeling and optimization in WSN context Hybrid model using real measurements on PowWow platform Static then dynamic energy optimization Transmit power and wake-up interval of sensor nodes MIMO Precoding Generalization of max-d min precoders Any number of data streams and constellation Implementation of precoded MIMO systems Cooperative communications Energy efficiency of new cooperation schemes Cooperative MIMO and relay, opportunistic communications Design of MAC protocols for cooperation HDR Defense, Olivier Berder, 10/12/12 6 / 42

10 HDR Defense, Olivier Berder, 10/12/12 7 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

11 HDR Defense, Olivier Berder, 10/12/12 7 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

12 Curriculum Vitae Education 1998 Maîtrise EEA, UBO 1999 DEA Électronique, option Signaux et Systèmes, UBO 2002 Doctorat en Électronique Professional activities 09/99-09/02 PhD Thesis at LEST/UBO Power allocation strategies and optimization of MIMO systems 10/02-09/04 ATER at LEST/UBO Automatic recognition of compression algorithms 10/04-01/05 Post-doc at France Telecom R&D Adaptive Rate-Distortion Optimised sound coder (ARDOR) Since 02/05 Associate Professor at ENSSAT (IRISA R2D2 then EPC CAIRN) HDR Defense, Olivier Berder, 10/12/12 8 / 42

13 Teaching Activities Preferred subjects Wireless Communications (EII3, LSI3, IMR2, ENI Gabès) Signal Processing (EII2, R&T2, UST Hanoi) Microprocessor, Architecture and Digital Systems (TC, EII1, LSI1, IMR1) Very useful to understand the whole panel of CAIRN activities HDR Defense, Olivier Berder, 10/12/12 9 / 42

14 HDR Defense, Olivier Berder, 10/12/12 10 / 42 Students Supervising A special thank to the Vietnamese connection!

15 Collaborative Projects Highlights Researcher Movies festival finals for CAPTIV FET Flagship Candidate Guardian Angels HDR Defense, Olivier Berder, 10/12/12 11 / 42

16 Publications Administration and expertise Elected member of IRISA Lab Council Reviewer for IEEE JSAC, TWC, TSP, TVT, TN, Globecom, ICC, etc. Program Committee of IEEE VTC, IWCLD General chair of WUPS2013 HDR Defense, Olivier Berder, 10/12/12 12 / 42

17 Publications Administration and expertise Elected member of IRISA Lab Council Reviewer for IEEE JSAC, T WC, SP, VT, N, IEEE Globecom, ICC, etc Program Committee of VTC, IWCLD General chair of WUPS2013 (deadline today!) HDR Defense, Olivier Berder, 10/12/12 12 / 42

18 HDR Defense, Olivier Berder, 10/12/12 13 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

19 HDR Defense, Olivier Berder, 10/12/12 14 / 42 WSN PowWow Platform Hardware components TI MSP430 Microprocessor TI CC2420 Radio transceiver Actel Igloo FPGA coprocessor Energy harvesting board Asynchronous MAC protocol: well suited to low traffic applications

20 Energy Modeling Scenario-based hybrid model [M. Cartron and M.M. Alam PhDs] Real-Time measurements for scenarios [CAIRN Development Team] Analytic expressions for trafic parameters Accurate energy consumption estimation [EURASIP JES 11] HDR Defense, Olivier Berder, 10/12/12 15 / 42

21 Energy optimization Transmit power optimization [M. Cartron PhD] Lifetime increase through static TX power tuning Different coding schemes investigated (implementation on low-power FPGA) Dynamic TX power adaptation Wake-up interval optimization [M.M. Alam PhD] Static adaptation to application constraints [DASIP07] Traffic-Aware Dynamic MAC protocol [IEEE JETCAS 12, IWCLD11, BSN11] Definition of Traffic Status Registers Self-adaptive algorithm Significant reduction of idle listening HDR Defense, Olivier Berder, 10/12/12 16 / 42

22 HDR Defense, Olivier Berder, 10/12/12 17 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

23 HDR Defense, Olivier Berder, 10/12/12 18 / 42 MIMO Systems Definitions Symbol vector s = (s 1, s 2,.., s b ) T Channel matrix H = [h ij ] n T transmitters, n R receivers Space-Time Block Codes Alamouti ( scheme: s1 -s C = ) 2 s 2 s 1 OSTBC by Tarokh

24 MIMO Systems CSIT Acquisition Channel State Information at the Transmitter Reciprocity: reverse channel state information (open-loop) Feedback: sending the forward channel information back to the transmitter (closed-loop) HDR Defense, Olivier Berder, 10/12/12 18 / 42

25 HDR Defense, Olivier Berder, 10/12/12 18 / 42 MIMO Systems Optimization criterion: max-d min Maximizing the minimum Euclidean distance (on the received constellations) Problem of the precoder matrix determination Space of solutions - number of variables in the precoder Constellation size and number of antennas

26 MIMO precoding Generalization of max-d min precoder [Q.T. Ngo PhD] Rectangular M-QAM modulations [IEEE T SP 12, ICC10, ICC09] Restriction to two precoders expressions Use of input shaping DFT-matrix [EURASIP JASP 12] Three data stream MIMO systems [IEEE T WC 12] New parameterization of the virtual channel Reducing the number of neighbors [WCNC11, EUSIPCO11] Towards an efficient implementation of precoders CORDIC algorithms for SVD [G. Garnier Eng., GRETSI09] On the efficiency of Sphere Decoders for precoded MIMO systems [V.H. Nguyen, (WCNC13)] HDR Defense, Olivier Berder, 10/12/12 19 / 42

27 HDR Defense, Olivier Berder, 10/12/12 20 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

28 HDR Defense, Olivier Berder, 10/12/12 20 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

29 HDR Defense, Olivier Berder, 10/12/12 21 / 42 Cooperative strategies Which one to choose between Direct transmission: fast but energy consuming (when possible)

30 HDR Defense, Olivier Berder, 10/12/12 21 / 42 Cooperative strategies Which one to choose between Direct transmission: fast but energy consuming (when possible) Multi-hop: variable latency

31 HDR Defense, Olivier Berder, 10/12/12 21 / 42 Cooperative strategies Which one to choose between Direct transmission: fast but energy consuming (when possible) Multi-hop: variable latency Cooperative relay: simple, reliable

32 HDR Defense, Olivier Berder, 10/12/12 21 / 42 Cooperative strategies Which one to choose between Direct transmission: fast but energy consuming (when possible) Multi-hop: variable latency Cooperative relay: simple, reliable Cooperative MIMO: efficient but synchronization requirement and complex reception

33 HDR Defense, Olivier Berder, 10/12/12 21 / 42 Cooperative strategies Which one to choose between Direct transmission: fast but energy consuming (when possible) Multi-hop: variable latency Cooperative relay: simple, reliable Cooperative MIMO: efficient but synchronization requirement and complex reception Opportunistic relaying: reliable but variable latency

34 HDR Defense, Olivier Berder, 10/12/12 22 / 42 ITS Application Context Application constraints and network topology can drive cooperative scheme choice Infrastructure to Vehicle (I2V) Communications in CAPTIV 1 Cooperative MIMO well suited to crossroads configuration 1 Cooperative strategies for low Power wireless Transmissions between Infrastructure and Vehicles

35 HDR Defense, Olivier Berder, 10/12/12 23 / 42 Energy efficiency of cooperative MIMO [T.D. Nguyen PhD] Energy models (literature, transceiver characteristics) Cooperative MISO more energy efficient from 30 meters Cooperation at the receiver not really energy efficient

36 Cooperative transmitters desynchronization problems Transmitters desynchronization [ICC08] ISI generated by 4 nearest symbols Performance degradation acceptable until 0.2T s HDR Defense, Olivier Berder, 10/12/12 24 / 42

37 Cooperative transmitters desynchronization problems Multiple sampling [VTC08] Time offsets estimation 2 consecutive samplings Recover signals orthogonality for space-time recombination Transmitters desynchronization [ICC08] ISI generated by 4 nearest symbols Performance degradation acceptable until 0.2T s HDR Defense, Olivier Berder, 10/12/12 24 / 42

38 HDR Defense, Olivier Berder, 10/12/12 25 / 42 Association of cooperative MIMO and relay techniques [L.Q.V. Tran PhD] Fully Distributed Space Time Codes [WCNC11]

39 Association of cooperative MIMO and relay techniques [L.Q.V. Tran PhD] Fully Distributed Space Time Codes [WCNC11] Exchange of data between relays before space-time transmission Useful for cooperative transmitters synchronization Low transmit power for transmission between relays HDR Defense, Olivier Berder, 10/12/12 25 / 42

40 Association of cooperative MIMO and relay techniques [L.Q.V. Tran PhD] Fully Distributed Space Time Codes [WCNC11] Exchange of data between relays before space-time transmission Useful for cooperative transmitters synchronization Low transmit power for transmission between relays HDR Defense, Olivier Berder, 10/12/12 25 / 42

41 Association of cooperative MIMO and relay techniques [L.Q.V. Tran PhD] Fully Distributed Space Time Codes [WCNC11] Exchange of data between relays before space-time transmission Useful for cooperative transmitters synchronization Low transmit power for transmission between relays HDR Defense, Olivier Berder, 10/12/12 25 / 42

42 HDR Defense, Olivier Berder, 10/12/12 26 / 42 Where should the relays be? r = dsr d sd rr = drr d sd fdstc always better than cdstc Performance decrease when relays far from each other

43 HDR Defense, Olivier Berder, 10/12/12 26 / 42 Where should the relays be? r = dsr d sd rr = drr d sd fdstc always better than cdstc Performance decrease when relays far from each other Significant SNR gain when relays further than middle distance

44 On the diversity-multiplexing trade-off of cooperative relays Diversity order increase when using fdstc Maximum diversity order for NR-relays (3) but not for R-Relays (2) DMT to find between all cooperative schemes HDR Defense, Olivier Berder, 10/12/12 27 / 42

45 Cooperative MAC Protocol Problems to deal with Nodes rendez-vous to exchange data Cooperative transmitters synchronization Quasi nothing in the literature (CL-MAC) ARQ-CRI MAC protocol [D.L. Nguyen Ms Thesis, (ICC13)] Automatic Repeat Request Cooperative Receiver Initiated Node waking up the soonest after destination becomes the relay Relaying only if no ACK from destination HDR Defense, Olivier Berder, 10/12/12 28 / 42

46 Cooperative MAC Protocol Problems to deal with Nodes rendez-vous to exchange data Cooperative transmitters synchronization Quasi nothing in the literature (CL-MAC) ARQ-CRI MAC protocol [D.L. Nguyen Ms Thesis, (ICC13)] Automatic Repeat Request Cooperative Receiver Initiated Node waking up the soonest after destination becomes the relay Relaying only if no ACK from destination HDR Defense, Olivier Berder, 10/12/12 28 / 42

47 OMNET++ Simulation results Network topology 5 source nodes, number of nodes in the relay region varies from 1 to 5 All nodes in both source and relay regions have a role of source nodes In cooperative networks, nodes in relay region can act as a relay HDR Defense, Olivier Berder, 10/12/12 29 / 42

48 OMNET++ Simulation results For low traffic, ARQ-CRI far more energy efficient For higher traffic, same energy but lower latency for ARQ-CRI For each optimal point, best energy-delay trade-off HDR Defense, Olivier Berder, 10/12/12 29 / 42

49 HDR Defense, Olivier Berder, 10/12/12 30 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

50 HDR Defense, Olivier Berder, 10/12/12 30 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

51 Scientific Outcome Supervising 4 defended PhD Thesis (+3 ongoing) 5 Ms Thesis (+2 ongoing) Publications 8 international journal papers (+3 submitted) 3 book chapters 22 international conference papers (+ 6 national) Collaborative Projects 2 ANR, 1 AMI ADEME, 1 Labex, 1 Equipex 1 ITEA (+ 1 FET Flagship) 1 PRIR + 2 EPT UEB 1 EA + 1 PICS HDR Defense, Olivier Berder, 10/12/12 31 / 42

52 MIMO precoding Various extensions of max-d min precoder Any rectangular QAM modulation Three data streams and possible combinations for more Reducing the number of neighbors Excellent results and good publication level But... First step towards implementation still unsufficient Enhance efforts to deal with the Agilent platform Lack of industrial and international collaborations in this domain HDR Defense, Olivier Berder, 10/12/12 32 / 42

53 Energy optimization of WSN Strong expertise in WSN Research & Development Link with other CAIRN team activities FPGA prototyping and low power design Protocol stack and software development Power-aware algorithms and power management Evaluated through analytics, network simulators or real platform PowWow Platform to promote Modular and evolutive hardware Open-source software Interest of industrial companies (e.g. Thales, CEA... ) HDR Defense, Olivier Berder, 10/12/12 33 / 42

54 Cooperative schemes Expertise in different strategies Cooperative MIMO ITS context for CAPTIV Cooperative relay Opportunistic relaying Need to compare them more accurately Cooperative MAC protocols Open subject in the literature Need for network simulation (WSNet, OMNET++/Mixim... ) Decrease the efficiency of cooperative techniques HDR Defense, Olivier Berder, 10/12/12 34 / 42

55 HDR Defense, Olivier Berder, 10/12/12 35 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

56 HDR Defense, Olivier Berder, 10/12/12 35 / 42 Outlines 1 WSN Context 2 Research Overview Teaching and supervising Collaborations and publications WSN Energy modeling and optimization MIMO Precoding 3 Cooperative techniques Energy efficiency of cooperative MIMO Cooperative relay protocols MAC protocols for cooperative relay 4 Summary 5 Perspectives

57 HDR Defense, Olivier Berder, 10/12/12 36 / 42 Towards a complete autonomy of wireless nodes Power manager design [GRECO, N.L. Trong PhD] Multi-source harvesting hardware Light, Heat, Moves, RF, Bio... Prediction algorithms Energy neutral operations Efficient implementation

58 Implementation of cooperative techniques This recent research domain lacks... experimental platforms! Proof of concept with real conditions Validating implementation problems such as: Synchronization needs (virtual MIMO for instance) Channel state information transmission in MH networks Self-tuning to find experimentally best strategies Equipex Future Internet of Things More than 50 software defined radio nodes (some MIMO compliant) AMI ADEME EGUISE Electric vehicles management Onboard multistandard radio nodes for communication HDR Defense, Olivier Berder, 10/12/12 37 / 42

59 HDR Defense, Olivier Berder, 10/12/12 38 / 42 Cooperative scheme selection

60 Cooperative scheme selection Given all possibilities above mentioned... When to cooperate? Whom to cooperate with? Which cooperative scheme to adopt? HDR Defense, Olivier Berder, 10/12/12 38 / 42

61 Cooperative scheme selection Given all possibilities above mentioned... When to cooperate? Whom to cooperate with? Which cooperative scheme to adopt? HDR Defense, Olivier Berder, 10/12/12 38 / 42

62 Cooperative scheme selection Given all possibilities above mentioned... When to cooperate? Whom to cooperate with? Which cooperative scheme to adopt? Complex (distributed) and open optimization problem that needs new paradigms HDR Defense, Olivier Berder, 10/12/12 38 / 42

63 Cooperative scheme selection Given all possibilities above mentioned... When to cooperate? Whom to cooperate with? Which cooperative scheme to adopt? Game theory algorithms [LDA, IEEE SPM 09] Nodes considered as players to reach stable operation points Coalitional game theory to model cooperative behavior Power allocation strategies Collaborative spectrum sensing Bio-inspired techniques [LR, ACM T AAS 12] Multi-agent smart systems Local decisions to reach global optimization HDR Defense, Olivier Berder, 10/12/12 38 / 42

64 Wireless Body Area Networks BAN cooperative techniques [V.H. Nguyen PhD] Power-aware adaptive coding in a distributed context Adaptive STBC coding scheme PhD cosupervised by C. Langlais (LAbSTICC) HDR Defense, Olivier Berder, 10/12/12 39 / 42

65 Wireless Body Area Networks BAN cooperative techniques [V.H. Nguyen PhD] Power-aware adaptive coding in a distributed context Adaptive STBC coding scheme PhD cosupervised by C. Langlais (LAbSTICC) Distributed MIMO precoding for BAN to base station transmission HDR Defense, Olivier Berder, 10/12/12 39 / 42

66 WSN Context Research Overview Cooperative techniques Summary Perspectives Wireless Body Area Networks BOdy World Interaction CominLabs Labex project 4 linked PhD subjects Cooperative communications Channel models Hardware architecture Algorithms for gesture recognition Guardian Angels for a smarter life EU Future and Emerging Technologies Flagship 10 year project HDR Defense, Olivier Berder, 10/12/12 40 / 42

67 HDR Defense, Olivier Berder, 10/12/12 41 / 42 Bibliographie I R. Zhang, O. Berder, J. Gorce, O. Sentieys, Energy-Delay Tradeoff in Wireless Multihop Networks with Unreliable Links, Ad Hoc Networks 10, 7, 2012, p Q. Ngo, O. Berder, P. Scalart, Minimum Euclidean Distance Based Precoders for MIMO Systems Using Rectangular QAM Modulations, IEEE Transactions on Signal Processing 60, 3, March 2012, p Q. Ngo, O. Berder, P. Scalart, Minimum Euclidean Distance Based Precoding for Three-Dimensional Multiple-Input Multiple-Ouput Spatial Multiplexing Systems, IEEE Transactions on Wireless Communications 11, 7, 2012, p M. Alam, O. Berder, D. Menard, O. Sentieys, TAD-MAC : Traffic-Aware Dynamic MAC Protocol for Wireless Body Area Sensor Networks, IEEE Journal on Emerging and Selected Topics in Circuits and Systems 2, 1, March 2012, p T. Nguyen, O. Berder, O. Sentieys, Energy-Efficient Cooperative Techniques for Infrastructure-to-Vehicle Communications, IEEE Transactions on Intelligent Transportation Systems, 12, 3, Sept. 2011, p M. M. Alam, O. Berder, D. Menard, T. Anger, O. Sentieys, A hybrid model for accurate energy analysis of WSN nodes, EURASIP Journal on Embedded Systems, Jan. 2011, p. 4 :1-4 :16.26 R. Zhang, J. Gorce, O. Berder, O. Sentieys, Lower Bound of Energy-Latency Trade-off of Opportunistic Routing in Multi-hop Networks, EURASIP Journal on Wireless Communications and Networking, L. Collin, O. Berder, P. Rostaing, G. Burel, Optimal minimum distance- based precoder for MIMO spatial multiplexing systems, IEEE Transactions on Signal Processing 52, 3, march 2004, p

68 HDR Defense, Olivier Berder, 10/12/12 42 / 42 Thanks for listening!