Telecommunication Networks Lab DET Department of Electronics and

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications A Finite State Modeling for Adaptive Modulation in Wireless OFDMA Systems Dania Marabissi, Daniele Tarchi, Federico Genovese, and Romano Fantacci University of Florence, Italy Email: tarchi@lenst. det. unifi. it 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 1

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Outline • Standard IEEE 802. 16 – S-OFDMA (Scalable-OFDMA) • • 11/04/2007 Adaptive modulation Proposed adaptation techniques Numerical results Conclusions COST 289 4 th Workshop - Gothenburg, Sweden 2

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Standard IEEE 802. 16 • Wi. MAX system, that is based on the IEEE 802. 16 specifications, offers broadband wireless accces in a wide are achieving about 50 Km in Lo. S connections and 10 Km in NLo. S connections • The standard aims to have until 70 Mbit/s per cell • Almost recently has been finalized the IEEE 802. 16 e version that adds the user mobility functionalities by exploiting the OFDMA technique • This allows to offers mutimedia services in mobility such as Internet, Voice over IP (Vo. IP) and video streaming 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 3

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications PHY characteristics • It transmits in the 2 – 11 GHz band – Recently has been almost defined four operational bands at 2. 3, 2. 5, 3. 5, 5 GHz • It allows to use QPSK, 16 QAM and 64 QAM, and variable coding rates • Both TDD and FDD (half and full duplex) are allowed • The channel width can varies from 1. 75 MHz to 20 MHz and in TDD the frame time can vary from 2 ms to 20 ms • The IEEE 802. 16 e has improved the OFDMA support 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 4

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications S-OFDMA • The IEEE 802. 16 e standard has introduced a particolar OFDMA called S-OFDMA (Scalable-OFDMA) • It mainly differs from the classical OFDMA because it can support several channel width by varying the FFT size during the subchannel mapping 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 5

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Adaptive Modulation • It selects automatically the modulation order for: – minimize the error probability; – maximize the transmitting rate; Channel state adaptativity • The AM (Adaptive Modulation) module manage the adaptivity based one: – channel state information and received power – receiver quality in terms of error rate 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 6

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications TDD Channel information • The proposed system is to be used in the TDD scheme • It aims to estimate the channel behaviour in the uplink and use the estimation for selecting the best modulation order to be used in the downlink 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 7

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Adaptive management Channel state adaptivity Three state Moore machine 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 8

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Maximum Throughput Technique ü Throughput: Thr(SNR, α, M) = log 2(M)[1 -SER(SNR, α, M)] ü For each fixed SNR we select two thresholds [αA, αB] by solving: Thr(SNR, αA, M=16) = Thr(SNR, αA, M= 4) Thr(SNR, αB, M=64) = Thr(SNR, αB, M=16) 64 QAM 16 QAM QPSK 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 9

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Minimum SER Technique ü For each fixed SNR at the BS two thresholds are selected [αA, αB]: SER(SNR, αA, M=16) = SER(SNR, M=4) SER(SNR, αB, M=64) = SER(SNR, M=4) where, supposing, 11/04/2007 SER(SNR, M=4) we have: COST 289 4 th Workshop - Gothenburg, Sweden 10

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Target SER Technique ü We can define a target SER value (TSER) based on the Quality of Service requested ü For each SNR we search two thresholds [αA, αB]: SER(SNR, αA, M=16) = TSER SER(SNR, αB, M=64) = TSER = constant SNR 11/04/2007 Example with SNR=15 d. B COST 289 4 th Workshop - Gothenburg, Sweden 11

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Decision thresholds 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 12

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications System Parameters • Channel model ITU-R M. 1225 (vehicular A) with 6 Rayleigh distributed main paths and Jake Doppler • Bandwidth equal to 10 MHz • NFFT = 1024 carriers • Carrier at 3. 5 GHz • Maximum speed equal to 125 Km/h • Frame duration TDD: – TF = 8 ms; – TDL=TUL=3, 995 ms; – (TTG=RTG=5μs) • OFDMA symbols in each frame: – 40 in DL + 39 in UL • Distributed subcarriers allocations 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 13

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Minimum SER and Maximum Throughput Maximum throughput allows For higher rate but SNR=8 d. B higher SER 16 QAM is the most used modulation order 2 d. B For SNR=16 d. B 16 QAM is the most used modulation order 11/04/2007 For SNR=16 d. B 64 QAM is the most used modulation order COST 289 4 th Workshop - Gothenburg, Sweden 14

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Target SER TSER is imposed based on the requested Qo. S level 11/04/2007 COST 289 4 th Workshop - Gothenburg, Sweden 15

Telecommunication Networks Lab. DET – Department of Electronics and Telecommunications Conclusions • • • 11/04/2007 IEEE 802. 16 e standard has been designed to perform a broadband wireless access by allowing multimedia services to mobile devices; this can be done thanks to the high flexibility in several parts of the system (like modulation, coding, subcarrier allocation). We propose a state model for adapt the modulation order to the channel behavior; three different techniques has been presented, that differs from the threshold calculation The Maximum throughput method is foreseen for real-time applications where throughput has more importance the error rate. The Minimum SER aims to achieve the best performance in terms of error probability for a certain SNR The Target SER aims to respect a target in terms of error probability in order to be considered to be used in specific applications. What we are working on: – – – Error rate as a performance indication; Extension of the model for considering variable coding rate jointly with modulation; Joint modulation/coding adaptation with subcarrier allocation, also by taking into account the upper layers queue size (opportunistic scheduling techniques). COST 289 4 th Workshop - Gothenburg, Sweden 16