Elham Torabi Supervisor: Dr. Robert Schober

Low-Rate Ultra-Wideband Low-Power for Wireless Personal Communication Area Networks Channel Models and Signaling Schemes Department of Electrical & Computer Engineering The University of British Columbia Elham Torabi Supervisor: Dr. Robert Schober

Outline 2 1. Overview and Introduction to the IEEE 802.15.4a Standard 2. Channel Models Generic Channel Model UWB Model Parametrization and Simulation Results for 2-10 GHz 3. Signaling Schemes Time-Hopping UWB (TH-UWB) RAKE Receivers UWB Transmitted-Reference (UWB-TR) UWB Differential Transmitted-Reference (UWB-DTR) Comparison

1. Overview and Introduction to the IEEE 802.15.4a Standard 3 The IEEE 802.15 low-rate alternative PHY task group 4a (TG4a) for WPANs, named subgroup IEEE 802.15.4a, has the mandate to develop an alternative physical layer for sensor networks and similar devices that work with the IEEE 802.15.4 MAC layer. Technical characteristics summary Topology Bit Rate Range Coexistence and Interference Resistance Power Consumption Quality of Service Antenna Complexity Location Awareness

1. Overview and Introduction to the IEEE 802.15.4a Standard 4 The principle interest of this subgroup is in providing communications for WPAN applications such as 1. Sensors networks 2. High precision positioning 3. Security/authentication 4. Smart home systems 5. Networks of wearable mobile devices 6. Real time location services

2. Channel Models 5 Different Proposed Channel Models in the IEEE 802.15.4a standard UWB channel models covering the frequency range from 2 to 10 GHz, considering indoor residential, indoor office, industrial, outdoor, and open outdoor environments (usually with a distinction between line-of-sight (LOS) and none-los (NLOS) properties) UWB channel model for the frequency range from 100 to 1000 MHz, considering a model for indoor office-type environments UWB channel model for the frequency range from 2 to 6 GHz, considering a model for body area networks (BANs) Main Goals are modeling the Attenuation Delay Dispersion

2. Generic Channel Model 6 Used for the 2-10 GHz frequency range. model treats only channel, while antenna effects should be modeled separately. block fading is assumed, i.e., channel stays constant over data burst duration. modified Saleh-Valenzuela (SV) model is adapted.

2. Generic Channel Model: Pathloss - Preliminary Comments 7 The pathloss for a narrowband system is conventionally defined as PL(d) = E {P RX(d, f c )}, P TX where P TX and P RX are transmit and receive power, respectively, d is the distance between transmitter and receiver, f c is the center frequency. Note that E { } = E lsf {E ssf { }}, where lsf and ssf indicate large-scale fading and small-scale fading, respectively. The pathloss related to wideband pathloss is defined as { } f+ f/2 ) 2 PL(f, d) =E H ( f, d d f, ) where H ( f, d connector. f f/2 is the transfer function from antenna connector to antenna

2. Generic Channel Model: Pathloss - Preliminary Comments 8 To simplify computations, we assume PL(f, d) =PL(f) PL(d) The frequency dependence of the pathloss is given as PL(f) f k, where k is the frequency dependence coefficient of the pathloss. The distance dependence of the pathloss in db is described by ( ) d PL(d) =PL 0 +10nlog 10, d 0 where the reference distance d 0 issetto1m,pl 0 is the pathloss at the reference distance, and n is the pathloss exponent.

2. Generic Channel Model: Pathloss - Recommended Model 9 According to the proposed model the pathloss is found to be PL(f,d) = P r (d, f) P TX amp (f) = 1 2 PL 0 η TX ant (f) η RX ant (f) ( ) 2(k+1) f f c ( ) n, d d 0 where P TX amp (f) is the output spectrum of the transmit amplifier, P r (d, f) is the received frequency-dependent power, η TX ant (f) is the frequency dependent transmit antenna efficiency, and η RX ant (f) is the frequency dependent receive antenna efficiency.

2. Generic Channel Model: Shadowing 10 Large-scale fading or shadowing is defined as the variation of the local mean around the pathloss, and has log-normal distribution about the mean. The pathloss, averaged over the small-scale fading in db, can be written as PL(d) =PL 0 +10n log 10 ( d d 0 ) + S, where S is a Gaussian-distributed random variable with zero mean and standard deviation σ S. If shadowing effects come into play, the overall channel is no longer wide sense stationary (WSS), therefore, for the simulation procedure according to the selection criteria document, shadowing shall not be taken into account.

2. Generic Channel Model: Power Delay Profile (PDP) 11 A statistical model for indoor multipath propagation is introduced, known as SV (Saleh-Valenzuela) model. The physical realization: received signal rays arrive in clusters. The cluster arrival times are modeled as a Poisson arrival process with some fixed rate Λ l. Subsequent rays arrive according to a Poisson process within each cluster, with another fixed rate. T l : arrival time of the lth cluster l =0, 1, 2,... τ k,l : arrival time of the kth ray measured from the beginning of the lth cluster k =0, 1, 2,... (aka Excess Delay)

2. Generic Channel Model: Power Delay Profile (PDP) 12 According to this model, the distribution of the cluster arrival times are given by a Poisson process p(t l T l-1 )=Λ l exp [ Λ l (T l T l-1 )], Ray arrival times are modeled with mixtures of two Poisson processes p(τ k,l τ (k-1),l )=βλ 1 exp [ λ 1 ( τk,l τ (k-1),l )] +(β 1) λ2 exp [ λ 2 ( τk,l τ (k-1),l ) where β is the mixture probability, λ 1 and λ 2 are the ray arrival rates. The number of clusters L is assumed to be Poisson-distributed ( L ( L) exp L) P L (L) = L!

2. Generic Channel Model: Pathloss - Preliminary Comments 13 The complex, low-pass impulse response of the channel L K h (t) = a k,l exp (jφ k,l ) δ (t T l τ k,l ), l=0 k=0 where a k,l is the gain of the kth ray of the lth cluster and the phases φ k,l are uniformly distributed in the interval [0, 2π]. The Power Delay Profile (PDP) of the channel is defined by taking the spatial average of h (t) 2 over a local area, in general P (t) K h (t) 2. For the SV model, and for the LOS case, the PDP, which is the mean power of the different paths, is found to be { E a k,l 2} 1 =Ω l γ l [(1 β) λ 1 + βλ 2 +1] exp ( τ k,l/γ l ),

2. Generic Channel Model: Power Delay Profile (PDP) 14 where Ω l is the integrated energy of the lth cluster,and γ l is the intra-cluster decay time constant. γ l k γ T l + γ 0, where k γ describes the increase of the decay constant with delay. k γ and γ 0 are intra-cluster decay time constant parameters. 10 log (Ω l ) = 10 log (exp ( T l /Γ)) + M cluster, where M cluster is a normally distributed variable with standard deviation σ cluster around it and Γ is the inter-cluster decay constant.

2. Generic Channel Model: Small-scale Fading 15 The distribution of the small-scale amplitudes a k,l, is Nakagami P X (x) = 2 ( m ) m x 2m 1 exp ( m ) Γ(m) Ω Ω x2 m 1/2 is the Nakagami m-factor, Γ(m) is the gamma function, and the parameter Ω corresponds to the mean power, and its delay dependence is thus given by the power delay profile. The m-parameter is modeled as a lognormally distributed random variable, whose logarithm has a mean µ m and standard deviation σ m. Both of these can have a delay dependence µ m (τ) =m 0 k m τ σ m (τ) = ˆm 0 ˆk m τ m 0 and k m are Nakagami-m factor mean and ˆm 0 and ˆk m are Nakagami-m factor variance.

2. Generic Channel Model: Auxilary Parameters 16 Mean Excess Delay: First moment of the PDP τ = P (τ) τdτ P (τ) dτ RMS Delay Spread: Square root of the second central moment of the PDP σ τ = τ 2 ( τ) 2 τ 2 = P (τ) τ 2 dτ P (τ) dτ

2. UWB Model Parametrization and Simulation Results for 2-10 GHz 17 Residential Environments: The model was extracted based on measurements that cover a range from 7-20 m, up to 10 GHz.

2. UWB Model Parametrization and Simulation Results for 2-10 GHz 18 Impulse responses for 100 realizations/channels.

2. UWB Model Parametrization and Simulation Results for 2-10 GHz 19 Mean excess delay.

2. UWB Model Parametrization and Simulation Results for 2-10 GHz 20 RMS delay spread.

2. UWB Model Parametrization and Simulation Results for 2-10 GHz 21 Average power delay profile (PDP).

2. UWB Model Parametrization and Simulation Results for 2-10 GHz 22 Number of significant paths within 10 db of peak.

2. UWB Model Parametrization and Simulation Results for 2-10 GHz 23 Number of significant paths capturing > 85% energy.

3. Signaling Schemes: Time-Hopping UWB 24 Time-Hopping UWB (TH-UWB) is a spread-spectrum technique used in impulse radio (IR) signaling and can be employed to support multiple-access scenarios. Pulse position modulation (PPM) and pulse amplitude modulation (PAM) are commonly used modulation schemes. A typical time-hopping format of the kth impulse radio transmitter output signal is ) ( ) (t (k) = w tr t (k) jt f c (k) T c d (k),k=0, 1, 2,...,K 1, s (k) tr j= where K is the number of transmitters, t (k) is the transmitter clock time, and w tr (t) represents the transmitted monocycle waveform. T f is the frame time or pulse repetition time. { } Each link uses a distinct pulse-shift pattern c (k) j called a TH sequence, which are pseudorandom with period N p. T c is the chip time. j j

The sequence 3. Signaling Schemes: Time-Hopping UWB 25 { d (k) j } j= is the data sequence, which is a sample sequence from a wide-sense stationary random process d (k) (t) A typical idealized monocycle.

3. Signaling Schemes: Time-Hopping UWB 26 Signal generated by the PPM-TH-UWB, in the case of {d j } =[1 0], T f =3e 9, N s =5, T c =1e 9, and TH Sequence= [10 102].

3. Signaling Schemes: Time-Hopping UWB 27 When K transmitters are active in the multiple-access system, the composite received signal r (t) at the receiver antenna is modeled as r (t) = K k=1 A k s (k) rec (t τ k )+n (t), where A k models the attenuation over the propagation path of the signal s rec (k) (t τ k ) received from the kth transmitter, the random variable τ k represents the time asynchronism between the clock of the signal received from transmitter k and the receiver clock. n (t) is White Gaussian noise.

3. Signaling Schemes: RAKE Receivers 28 System model. The zero-mean i.i.d. data symbols {d n } are passed through a unit energy pulse shaping filter w tr (t). After pulse shaping, the signal undergoes the effects of a channel with L paths whose response given by h (t) = L 1 l=0 α l δ (t τ l ), where α l and τ l are the attenuation and delay introduced by the lth path of the channel.

3. Signaling Schemes: RAKE Receivers 29 The received signal can then be expressed as L 1 r (t) = d i α l w tr (t τ l it f )+w(t), j= l=0 In the case of no ISI and when the noise is AWGN, the optimal receiver is a filter matched to the received waveform, this is implemented in a RAKE receiver structure with M arms, which can be represented as a filter with response f (t) = M 1 m=0 c m w tr ( t φ m ), where c m is RAKE tap coefficient, and φ m is RAKE delay.

3. Signaling Schemes: RAKE Receivers 30 The sampled output of the RAKE receiver is then y n =[r (t) f (t)] t=ntc = i= L 1 d i l=0 M 1 m=0 α l c m R wtr (nt c it c + φ m τ l )+ w (n where R wtr (t) w tr (τ) w tr (τ + t) dτ is the time-autocorrelation of the pulse shape, and w (t) is filtered noise. The optimal combiner for the AWGN multipath channel is maximum ratio combining (MRC), wherem = L fingers, c m = α m,andφ m = τ m,andwhen the received signals on each finger are orthogonal (as is the case when there is no ISI), MRC attains the matched filter bound.

3. Signaling Schemes: UWB Transmitted-Reference (UWB-TR) 31 The transmitted signal of a UWB transmitted-reference (UWB-TR) system with antipodal modulation is given by [ s tr (t) = wtr (t it f )+b i/ns w tr (t it f T d ) ], i= where w tr (t) is a transmitted monocycle waveform, and T f is its pulse repetition or frame time. Each bit is transmitted in N s successive frames to achieve an adequate bit energy in the receiver. The data bits are b i/ns {+1, 1} with equal probability.

3. Signaling Schemes: UWB Transmitted-Reference (UWB-TR) 32 The received TR signal in a stationary channel over a bit time is modeled by r (t) = N s 1 L [ αl w rx (t it f τ l )+α l b i/ns w rx (t it f T d τ l ) ] +n (t), i=0 l=1 where L is the number of specular propagation paths, lth path s propagation delay and amplitude are being denoted by τ l and α l,andn (t) is white Gaussian noise. There are a few types of receivers for TR signals; Generalized likelihood ratio test (GLRT) receiver, simple transmitted reference (STR) receiver, and averaged transmitted reference (ATR) receiver.

3. Signaling Schemes: UWB Transmitted-Reference (UWB-TR) 33 GLRT receiver BEP: [ P bit = Q 2 N s ( N0 E f ) + L N 2 s ( N0 E f ) 2 ] 1 2, STR receiver BEP: [ P bit = Q 2 N s ( N0 E f ) + 2WT mds N s ( N0 E f ) 2 ] 1 2, ATR receiver BEP: [ P bit = Q 2 N s ( N0 E f ) + 2WT mds N 2 s ( N0 E f ) 2 ] 1 2, where W is the one-sided noise bandwidth of the receiver, E f is the received energy per frame at two pulses per frame, and T mds = τ l + T w.

3. Signaling Schemes: UWB Differential Transmitted-Reference (UWB-DTR) 34 In UWB Differential Transmitted-Reference (UWB-DTR) system no references are transmitted, but instead, the data signal in the previous frame is used as reference. The modulation and demodulation techniques of this DTR system are similar to those used in differential phase shift keying (DPSK). The differentially modulated UWB transmitted signal is s tr (t) = m i w tr (t it f ), i= where m i = m i 1 b i/n s, and all the other parameters are the same as defined for the TR system.

3. Signaling Schemes: UWB Differential Transmitted-Reference (UWB-DTR) 35 Block diagram of the modulator and demodulator in DTR UWB system.

3. Signaling Schemes: UWB Differential Transmitted-Reference (UWB-DTR) 36 The received signal of this differential system is L r (t) = α l m i 1 b i/ns w rx (t it f τ l )+n(t) i= l=1 DTR receiver BEP: [ P bit = Q 2N s 1 Ns 2 ( N0 E p ) + WT mds 2N s ( N0 E p ) 2 ] 1 2, where W is the one-sided noise bandwidth of the receiver, E p is the received energy per pulse at one pulse per frame, and T mds = τ l + T w.

3. Signaling Schemes: UWB Differential Transmitted-Reference (UWB-DTR) 37 BEP of GLRT, ATR, DTR and STR receiver structures in a dense resolvable multipath environment.

3. Signaling Schemes: Comparison 38 One of the challenges of a UWB system implementation is providing very stable reference clocks for the transmitter and receiver pulse repetition frequency (PRF) generators. The differential phase shift keying (DPSK) receiver, which is used to detect the DTR modulated UWB signals (also known as differential detector), is much less sensitive to jitter on the receiver PRF clock than the RAKE receiver.

3. Signaling Schemes: Comparison 39 Sensitivity of RAKE and DPSK receivers to PRF clock time offset.

3. Signaling Schemes: Comparison 40 Another challenge for the RAKE receiver is to generate an impulse that closely matches the received impulse at the input to the receiver. Since the DPSK receiver is correlating with the delayed replica of itself (although noisy), the distortions will automatically be accounted for as long as the channel is slow relative to the PRF, which will typically be the case.

3. Signaling Schemes: Comparison 41 The required delay needed by the RAKE receiver for each arm is unknown at the beginning of a communications session and it must be found using some type of search procedure. The DPSK receiver does not have this search requirement, since it continuously correlates the received signal with a delayed replica. As a result, the DPSK receiver architecture could be used to rapidly acquire the transmitted packets without a long training sequence or search algorithm.

3. Signaling Schemes: Comparison 42 For a low-power low data-rate UWB system, the DPSK receiver architecture could be a simple and low-cost alternative to the traditional RAKE receiver.

43 Thank you!