Proposal of a multi-standard transceiver for the WBAN Internet of Things

Transcription

1 Proposa of a muti-standard transceiver for the WBA Internet of Things Jianxiao Yang, Benoit Geer, Tarak Arbi To cite this version: Jianxiao Yang, Benoit Geer, Tarak Arbi. Proposa of a muti-standard transceiver for the WBA Internet of Things. ISIVC,2016 Internationa Symposium on Signa, Image, Video and Communications, ov 2016, Tunis, Tunisia. pp , 2016, Signa, Image, Video and Communications (ISIVC), Internationa Symposium on. < /ISIVC >. <ha > HAL Id: ha Submitted on 30 Oct 2017 HAL is a muti-discipinary open access archive for the deposit and dissemination of scientific research documents, whether they are pubished or not. The documents may come from teaching and research institutions in France or abroad, or from pubic or private research centers. L archive ouverte puridiscipinaire HAL, est destinée au dépôt et à a diffusion de documents scientifiques de niveau recherche, pubiés ou non, émanant des étabissements d enseignement et de recherche français ou étrangers, des aboratoires pubics ou privés.

2 PROPOSAL OF A MULTI-STADARD ASCEIVER FOR THE WBA ITERET OF THIGS Jianxiao Yang, Benoit Geer, Tarak Arbi Department U2IS, ESTA-ParisTech, Paaiseau, France jianxiao.yang,benoit.geer,tarak.arbi@ensta-paristech.fr ABSACT In this paper, an origina physica ayer design of the IEEE narrowband receiver is proposed, incuding frame synchronization, timing synchronization, and carrier frequency synchronization. This proposed singe-carrier system can be incorporated into a conventiona muti-carrier system of the IEEE transceiver by reusing its hardware resource. Simuation resuts show that the performances of the proposed system approach cosey to the theoretica performances. Index Terms Wireess Body-Area etwork (WBA), Internet of Things (IoT), Physica Layer Design, IEEE , Frame Synchronization, Timing Synchronization, Carrier Frequency Synchronization, IODUCTIO The IEEE standard [1] is a short-range (within 3 m range), ow power, wireess communication standard in cose proximity to body area to offer reiabe rea-time medica and non-medica services. Medica WBAs (Wireess Body Area etwork) consist of impant devices and wearabe medica systems to measure the heath status. on-medica WBA incudes wearabe consumer eectronics and entertainment devices. Differenty from the UWB mode of the IEEE standard targeting nonmedica appications, the narrowband mode standard [2] particuary aims at becoming an IoT (internet of things) standard for the future heathcare and medica industries. Therefore, this paper focuses on the narrowband mode. Over the 2.4-GHz ISM band, severa important wireess communication standards coexist, incuding IEEE 02.11a/b/g/n (WiFi) [3], IEEE (Buetooth) [4], IEEE (ZigBee) [5], and IEEE (Wireess Body-Area etwork, WBA) standards. Among these standards, the IEEE 02.11a/b/g/n standard is the most popuar wireess communication standard. For the WBA standard, it woud be extremey attractive to have the same coverage as the WiFi standard. Therefore, differenty from the works in [6]-[12], this paper intends at designing a singe-carrier receiver for the IEEE narrowband mode by expoiting muti-carrier signa processing agorithms, such as used in the 02.11a/b/g/n standard. This paper is organized as foows. The physica ayer specifications of the IEEE narrowband signa are recaed in section 2. The WiFi compatibe designs for the physica ayer of IEEE standard are detaied in section 3. The performances of the proposa are dispayed in section 4. Concusion is made at the end of this paper. 2. PHYSICAL LAYER SPECIFICATIO OF IEEE ARROWBAD SIGAL The IEEE physica ayer frame structure is shown in Fig.1. The physica ayer consists of three parts, i.e., the Physica Layer Convergence Protoco (PLCP) preambe, the PLCP header and the physica ayer service data unit (PSDU). The atter two parts together are caed Physica ayer Protoco Data Unit (PPDU). 90 Bits PLCP Preambe Rate Reserved Length PLCP Header (31 Bits) PHY Header 15 Bits Reserved PPDU Burst Scramber Mode Seed 4 Bits 12 Bits HCS 56 Bits PSDU Bits BCH MAC MAC Parity Header Frame Body 16 Bits FCS Fig.1. IEEE WBA physica ayer frame structure The first part is the PLCP preambe. It serves as a known training sequence for baseband processing functions, such as frame synchronization (or packet detection), timing synchronization, frequency synchronization. For the narrowband mode, two types of PLCP preambes are defined in order to reduce fase aarms coming from the networks operating on adjacent channes. Each preambe is defined by concatenating a 63-bit M-sequence with the foowing 27-bit extension sequence Since the preambe bit sequence is 2 -DBPSK moduated, the preambe contains 90 compex-vaued os. The second part of the frame is the PLCP header and is aso the first component of the PPDU. The 31-bit-PLCP header incudes the most important physica ayer coding and moduation parameters, a header check sequence (HCS) and BCH parity bits. Since this part is aso 2 -DBPSK

3 moduated, the PLCP header contains 31 compex-vaued os. The ast part of the frame is the PSDU is aso the ast part of the PPDU. This part contains the MAC header, a MAC frame body which contains the usefu information and a frame check sequence (FCS). The PSDU is then encoded and spread/intereaved before being scrambed. For the narrowband mode, the binary sequence in the PSDU is mapped onto one of the three rotated and differentiay encoded consteations: -DBPSK, -DQPSK or - DPSK. b 2k b 2k 2 Tabe I. bk, 0 1 Tabe II. b 2k 1 0, 0 0, DBPSK k 2 2 -DQPSK k , , 1 Tabe III., b2k 1, b3k 2 0, 0, DPSK k 0, 0, 1 3 0, 1, 0 7 0, 1, 1 5 1, 0, 0 1, 0, 1 1, 1, 0 1, 1, For any of the previous differentiay encoded consteations, the binary bit stream bn, n = 0, 1,, 1 is mapped onto a corresponding compex-vaued sequence sk, beow: k = 0, 1,, 2 og M 1 as shown as sk sk 1exp j k, (1) where M is the consteation order, s1 exp j 2 is the reference for the first o of the preambe and the phase transitions between os are given in Tabe I, Tabe II or Tabe III for 2 -DBPSK, 4 -DQPSK or -DPSK, respectivey. Due to the short communication range, as we as the ow o rate, the communication channe can be treated as a k fat-fading and sow-varying channe. Therefore, the received observation signa can be modeed as: rt h snut nt wt, (2) where n ut is the conventiona square-root raised cosine (SRRC) puse shaping fiter, is the sow-varying fatfading channe attenuation and is treated as a constant within a packet reception period. Based on (2), the received baseband processing shoud incude frame synchronization, carrier frequency synchronization, and timing synchronization before performing differentia DMPSK o detection. It shoud be mentioned that carrier phase synchronization is not performed thanks to the differentia detection appied for demapping DMPSK os. 3. PHYSICAL LAYER DESIG In this section, the various baseband signa processing bocks frame synchronization, carrier frequency synchronization and timing synchronization are detaied Frame synchronization Since the IEEE system is a packet-based transmission system, ocating the start of frame (SOF) shoud be the first task to be estabished before any other baseband processing. Therefore, the frame synchronization shoud toerate various front-end impairments such as ing cock timing offset, carrier frequency offset, and additive noise. In order to perform robust frame synchronization over the received es with ing rate f 2 f, the foowing time-domain correation agorithm is proposed: nˆ SOF L 9 T arg max r n p r n p exp j pq, n 1 p0 T q0 (3) where is the ˆn SOF h is the estimated start of frame (SOF), r n received overed observation at time instante, T 1 f is the o time interva, T 1 f is the ing time interva, is the n-th phase transition between the successive DMPSK os sn and sn 1. Two points shoud be noted: First, the differentia T correation term r n r n reduces the impact of the T carrier frequency offset without having been estabished yet. Second, the index indicates that the correation is not ony performed between consecutive o-time spaced es, it is aso performed between with severa otime spaced es. This improves the robustness of the frame synchronization. n n

4 In the proposed frame synchronization, is enough to ocate SOF with a carrier frequency offset f 0.1 f and a signa-to-noise ratio SR = -1 db. L Carrier frequency synchronization Proceeded the frame synchronization, the carrier frequency synchronization is the second to be estabished. It has huge impact over the timing synchronization and the o detection. Simiary to a conventiona OFDM system, the carrier frequency synchronization consists of a coarse (carrier frequency offset) synchronization step and a fine synchronization step. The coarse synchronization is performed over the received overed observations as foows: 1) Transform every overed time-domain preambe observations into the frequency domain by. The corresponding frequency-domain preambe observations are designated as, = 0, 1,, 1 Rk 2) Transform the corresponding time-domain preambe into the frequency domain by. The corresponding frequency-domain preambe observations are designated as, = 0, 1,, 2 1 P k 3) Perform shift correation and ocate the maximum correation between the two frequency-domain sequences as: nˆ coarse 41 R p nmod P p p0 arg max, n 21 R p nmod P p p 4 2 (4) where mod means a modue operation. The estimated fˆ f coarse 2 is then equa to: f nˆ (5) coarse coarse 4) Compensate the f coarse. It can be noted that the estimated by the coarse estimation agorithm has a precision of f. In this agorithm, the size arger the vaue k k determines the precision: the, the finer the precision of the estimator. In the proposed coarse estimator, = 12 is used corresponding to 12 preambe es and corresponding to 64 preambe os due to f 2 f. Therefore, for every packet containing 90 preambe os, there are two partiay correated estimates f coarse. In the proposed system, with = 12, the residua can aways be suppressed within the order of for an initia f carrier frequency offset f 0.1 f and a signa-to-noise ratio SR = -1 db. After the coarse compensation, a fine synchronization is performed over the coarse compensated time-domain preambe os as foows: ˆ 1 f fine 2 LT r n (6) L tan r p r p exp j pq p0 q0 In order to simpy the system compexity, the fine is estimated as foows: where ˆ v fine fˆ fine 1 1 tan vˆ fine, (7) T is computed by using a one-tap IIR fiter: 1 vˆ 1 vˆ r p r p exp j. () fine fine p q q0 In the proposed fine estimator, the vaues 0.01 are used. L 2 and 3.3. Timing synchronization In the proposed design, the timing synchronization is the ast task to be achieved before the differentia o detection. Its objective is to estimate the deay parameter in the signa rt, where one has n (9) n. r t h s u t nt w t Differenty from a conventiona timing synchronization is generay composed of two bocks, namey the o interpoator and the timing error detector (TED), the proposed timing synchronizer intends to estimate the goba hu t nt and behaves channe impuse response more ike a channe estimator. Therefore, the timing recovery invoves two different bocks: channe estimator and channe equaizer. The channe estimator aims at estimating the frequencydomain channe transfer function based on the compensated overed observations and the known preambe os, on thus performs the foowing steps: 1) Transform every overe time-domain preambe observations into the frequency domain by. The corresponding frequency-domain compensated preambe observations are designated as Rk, k = 0, 1,, 1 2) Insert 1 zero for every preambe o in the time domain, and then transform the zero-inserted preambe os into the frequency domain by.

5 The corresponding frequency-domain preambe piots are designated as Pk, k = 0, 1,, 1 3) Like for a conventiona OFDM channe estimator, the channe transfer function can be estimated as Ĥ k Rk Pk 4) Linear interpoate the frequency-domain channe transfer responses between consecutive Ĥk vaues. One has a new interpoated frequency-domain channe transfer function of ength 2. Hˆ I k The channe equaizer then aunches the overap-save (OLS) method [13] to perform the frequency-domain equaization over the twice-overed es as foowing: 1) Take the ast overed observations from the previous OLS bock 2) Take the newy overed observations which need to be equaized 3) Transfer the 2 es into the frequency domain which are then designated as Rk 4) Perform a inear equaization as D ˆ k Rk Hˆ I k 5) Transform the equaized frequency-domain es back into the time domain ˆD k 6) Output the ast es, among which, the first e of every two es is the timing recovered DMPSK o. In the proposed timing synchronizer, 64 is used for the channe estimation. 4. SIMULATIO RESULTS In this section, the performance of the proposed system for -DBPSK (see Fig.2), -DQPSK (see Fig.3) and - DPSK (see Fig.4) is evauated for two scenarios: one is the simuation with a random timing offset (RTO) T 2, T 2 (see the curves indicated as RTO Sim ) and the other is with a RTO and a arge f f (see the curves indicated as RTO + 10% Sim ). In these simuations, a the baseband processing bocks (frame synchronization, carrier frequency synchronization and timing synchronization) are active and therefore the estimation errors exist. In order to be fairy compared with the corresponding theoretica performances for differentia detected 2 -DBPSK, 4 -DQPSK and -DPSK signas, the simuated performances are evauated without the use of the BCH decoder. Moreover, the theoretica performance (indicated as Theory ) represents the best performance over AWG channe with differentia detection method and serve as a reference performance [14]. BER SR (db) Fig.2. The simuation performance of the proposed system for a -DBPSK signa. BER 2 SR (db) Fig.3. The simuation performance of the proposed system for a -DQPSK signa. BER /2 DBPSK /4 DQPSK RTO Sim Theory RTO + 10% Sim RTO Sim Theory RTO + 10% Sim / DPSK RTO Sim Theory RTO + 10% Sim SR (db) Fig.4. The simuation performance of the proposed system for a -DPSK signa. It can be observed from Fig.2 to Fig.4 that the bit error rate (BER) curves foow cosey the theoretica performance for

6 a the 2 -DBPSK, 4 -DQPSK and -DPSK signas with ony RTO. As the increases, the BER curves deviate from the corresponding theoretica curves as expected from the nature of the differentia moduation. As the consteation order increases, the signa becomes more and more sensitive to the. 5. COCLUSIO In this paper, the design of physica ayer agorithms for IEEE narrowband receiver was presented. Differenty from the conventiona design for a singe-carrier transceiver, the proposed signa processing agorithms can directy be appied to an OFDM based communication system such as WiFi [3]. Through numerica simuations, the proposed agorithms foow cosey the corresponding theoretica performances over an AWG channe with random timing errors. This indicates that the proposa introduces a negigibe eve of estimation error. This design is therefore attractive for the design of a muti-standard and muti-mode transceiver and can potentiay push the IEEE WBA standard to more appications. In the future we aim at incuding in our muti-standard receiver a soft decoder [15]-[16] to both take fuy advantage of a soft syn chronization [17]-[1] and of a cross-ayer design [19]-[23]. Acknowedgement : This study was funded in part by the European project H2020 Bridges. 6. REFERECES [1] IEEE Standard for Loca and metropoitan area networks Part 15.6: Wireess Body Area etworks, [2] C. Lee, J. Kim, H. S. Lee and J. Kim, "Physica ayer designs for WBA systems in IEEE proposas," in the proceedings of 9th Internationa Symposium on Communications and Information Technoogy, ISCIT pp , Icheon, [3] IEEE Standard for Loca and metropoitan area networks Part 11: Wireess LA Medium Access Contro (MAC) and Physica Layer (PHY) Specifications, [4] IEEE Standard for Loca and metropoitan area networks Part 15.1: Wireess Medium Access Contro (MAC) and Physica Layer (PHY) Specifications for Wireess Persona Area etworks (WPAs), [5] IEEE Standard for Loca and metropoitan area networks Part 15.4: Low-Rate Wireess Persona Area etworks (LR- WPAs), [6] B. Choi, B. Kim, S. Lee, K. Wang, Y. Kim and D. Chung, "arrowband Physica Layer Design for WBA System," in the proceedings of First Internationa Conference on Pervasive Computing Signa Processing and Appications (PCSPA), 2010, pp , Harbin, [7] BoHeun Choi, ByungSoo Kim and SangSeo Lee, et a. "arrowband Physica Layer design for WBA system," in the proceedings of First Internationa Conference on Pervasive Computing, Signa Processing and Appications (PCSPA), pp , Sep [] Mengyuan Chen, Jun Han, Dabin Fang, Yao Zou and Xiaoyang Zeng, "An Utra Low-Power and Area-Efficient Baseband Processor for WBA Transmitter," in the proceedings of Signa and Information Processing Association Annua Summit and Conference (APSIPA), [9] Yunping Liang, Yu Zhou and Ye Lu, "The design and impementation of IEEE Baseband on FPGA," in the proceedings of Internationa Conference on Heath Informatics, IFMBE Proceedings 42, ov [10] P. Mathew, L. Augustine, D. Kushwaha, V. Desaphine and A. David Sevakumar, "Impementation of B PHY transceiver of IEEE WBA on FPGA," in the proceedings of 2015 Internationa Conference on VLSI Systems, Architecture, Technoogy and Appications (VLSI-SATA), pp. 1-6., Bangaore, [11] P. Mathew, L. Augustine, D. Kushwaha, D. Vivian and D. Sevakumar, "Hardware impementation of B PHY baseband transceiver for IEEE WBA," in the proceedings of 2014 Internationa Conference on Medica Imaging, m-heath and Emerging Communication Systems (MedCom), pp , Greater oida, [12] G. Devita et a., "A 5mW muti-standard Buetooth LE/IEEE SoC for WBA appications," in the proceedings of European Soid State Circuits Conference (ESSCIRC), ESSCIRC th, pp , Venice Lido, [13] Proakis, J. G., and Manoakis, D. K., Digita Signa Processing, 4th Edition., Pearson, [14] Xiong, F., Digita Moduation Techniques (Second Edition), Artech House, Apri, [15] I. Diatta, D. De Geest, B. Geer, "Reed Soomon Turbo Codes for High Data Rate Transmission", Proceedings of IEEE VTC, pp , Mian, May [16] B. Geer, I. Diatta, J.P. Barbot, C. Vanstraceee, F. Rambeau, " Bock Turbo Codes : From Architecture to Appication", Proceedings of IEEE ISIT, Seatte, Juy [17] J. Yang, B. Geer, and A. Wei, "Bayesian and Hybrid Cramer-Rao Bounds for QAM Dynamica Phase Estimation", Proceedings of IEEE ICASSP'09, Taipei, Apri [1] J. Yang, B. Geer, C. Herzet, and J.M. Brossier,"Smoothing PLLs for QAM Dynamica Phase Estimation", Proceedings of IEEE ICC'09, Dresden, June [19] L. Zhou, B. Geer, X. Wang, A Wei, B. Zheng, H.C. Chieh, "Muti-user video streaming over mutipe heterogeneous wireess networks: a distributed, cross-ayer design paradigm", Journa of Internet, vo. 10, no 1, [20] L. Zhou, B. Zheng, J. Cui, B. Geer, A.Wei, S. Xu, "Cross- Layer Design for Fow Contro in Cooperative Muti-Hop Wireess etworks", Internationa Journa of Innovative Computing, Information and Contro, vo. 5 no 2, Feb [21] L. Zhou, X. Wang, Y. Li, B. Zheng, B. Geer, "Optima scheduing for Mutipe Description Video Streams in Wireess Muti-Hop etworks", IEEE Communications etters, vo. 13,no.7, pp. 1-3, Juy [22] L. Zhou, B. Geer, B. Zheng, J. Cui, " Cross-Layer Design for Scheduing in Cooperative VAETS", Proceedings of IEEE ITST,Lie, October [23] L. Zhou, B. Geer, A. Wei, B. Zheng, J. Cui, S. Xu, "Cross- Layer Rate Aocation for Mutimedia Appications in Pervasive Computing Environment", Proceedings of IEEE GLOBECOM 200, ew Oreans, ov. 200.