ece BRADLEY DEPARTMENT of ELECTRICAL & COMPUTER ENGINEERING

Transcription

1 Reza Rezaiesarlak Majid Manteghi November 24

2 Outline Radio Frequency Identification Systems Chipless RFID system Tag Design Complex natural resonance-based design of chipless RFID tags Design of chipless tag using the theory of characteristic modes Detection Technique Short-time matrix pencil method (STMPM). Identification of the tag. Localization and collision avoidance algorithm.

3 Radio Frequency Identification Systems. Punched card 2. Barcode D barcode Basile Bouchon s punched paper tape system Musée des Arts et Métiers, Paris. 2D barcode

4 Radio Frequency Identification Systems 3. Near-Field Magnetic Tags 4. Chipless RFID Tags 5. SAW Tags Piezoelectric Substrate Multi-Turn Coil

5 Radio Frequency Identification Systems 6. UHF On-Chip Tags Tag Antenna Interrogation impulse Demodulator Reader Backscattered signal from the tag Voltage Multiplier Digital Logic Rived signal by the reader Modulator Integrated Circuit Gamma Matching These arms can adjust to reduce the antenna length A capacitive load allows miniaturization

6 Chipless RFID System UWB Antenna Reader UWB Antenna E (v/m) x RCS (dbv/m) time (ns) frequency (GHz)

7 Reader Architecture TX Antenna Chirp Generator PA Directional Coupler Digital Baseband: DSP Demodulator Control Signals Data Acquisition Digital Baseband: IO Networking DSP Control Signals Protocol Data Acquisition Networking DC Cancellation ADC AGC ADC DC Cancellation Delay Line I 9 Q t LO RX Antenna LNA

8 RFID tags 25, Jalaly 26, Engheta 27, Manteghi 29, Karmakar 22, Costa 23, VTAG

9 Tag Design Chipless RFID tag acts as both encoder and scatterer. Design parameters Resonant frequency Quality-factor (Damping factor) of the resonator Power of the CNR s j RCS of the tag E Q P stored radiation

10 Singularity Expansion method (SEM) v ( t) h( t) v ( t) R V ( s) H( s) V ( s) R Based on SEM: N R H s s s n E() s T nn n T v R N v T (t) v R (t) j n jn t t Re R e et nn n Late-time Entire-domain function Complex natural resonances (CNRs): s n n j n

11 E (V/m) E (dbv/m) Singularity Expansion method (SEM) -3-4 x -5 y T=.2ns T=.4ns Frequency (GHz).2 Frequency-domain response. -. T=.4ns T=.6ns Time (ns) Time-domain response

12 R f (GHz) Complex Natural Resonance-based design of chipless RFID Tag s W a L d=.8mm d=.4mm (s - ) -.6 x x y d L (mm) W (mm) x (s - ) a=2mm a=mm a=8mm d (mm) incident angle (degree)

13 Design of tag based on theory of characteristic modes EFIE: G ~ ( r, r; s ), J( r; s) tˆ E inc r; G Z R jx Weighted eigenvalue equation: X,, Jn n R J n s E Chipless tag The characteristic modes are orthogonal. The eigenvalues and characteristic modes are real values. At the resonant frequency of the structure, n The radiated fields from the characteristic modes in the far-zone region of the scatterer are orthogonal. Current distribution: Modal significance: Characteristic angle: inc J n, E J n 8 tan MSn j J n jn n n n

14 Eigenvalue Modal significance Design of tag based on theory of characteristic modes Mode Mode frequency (GHz) Mode Mode frequency (GHz) d s mm d s. 3mm

15 RCS (dbm 2 ) 24-bit Tag and its RCS Chipless RFID tag -35 ID ID Frequency (GHz) Radar Cross Section of the tag

16 E (dbv/m) group delay (sec) Detection technique Frequency-domain techniques Absolut value of Backscattered field Group-delay x -6 y z x -8 =45, =9 =45, =85 =2, = frequency (GHz) =45, =9 =75, =-4 =, = frequency (GHz) 3-bit tag Electric field Group delay

17 Detection technique Time-frequency techniques Short-time Fourier transformation (STFT) Wavelet transform WVD (Wigner-Ville Distribution) Time-frequency distribution series Windowed super-resolution algorithm Adaptive Gaussian representation Joint Reassigned Time-Frequency method (JRTF) Joint time-frequency based on compressed sensing

18 Amplitude (V/m) Example (Time-Frequency Analysis of the signal) s ( t) 3e 3e t t sin sin 2f t 2t 2f t 2e sin2f t 2 t t t t Time (ns) Signal in time domain

19 Frequency (GHz) Frequency (GHz) STFT Analysis Time (ns) T=5ns Time (ns) T=.64ns

20 s Wavelet Analysis x

21 E (mv/m) Short-time matrix pencil method (STMPM) Time-domain signal: N E s r, τ = Re R n e s n τ τ n U(τ τ n n= Windowed signal: N E s r, τ = Re R n T e s n τ τ n u(τ τ n n= + e(τ).5 T W Residue: R n T = R n e s nt = R n e (α n+jω n )T In logarithmic scale: Ln R n T = Ln R n α n T -.5 T Time (ns)

22 T (ns) f (GHz) T (ns) T (ns) STMPM (Results) 3 Time-Frequency 3 Time-damping factor 2.5 P P2 P3 2.5 P P3 P turn-on time=.65 ns frequency (GHz) P3 Time-Residue P Ln( R n ) P x Pole diagram x 9 P P2 P3

23 Time (ns) Time (ns) E (V/m) STMPM vs STFT x Cylinder excited by an plane wave time (ns) Time domain response Frequency (GHz) From STMPM 5 5 Frequency (GHz) From RJTF From STFT

24 Early-time response (Altes model) Scattered field E( r; s) s I G k Early-time response e e ( r; t) M m A m p( t t r, r, J( r; ) 2 s m ) Transmitted signal P(t) A m ( p( t tm dp( t) ) p( t tm )) dt t t m t 2 A A 2 Tapped Delay Line t AM t M p( t) p( t d) p( t kd) kd p( t) dt Adder Scattered signal

25 Altes model Early-time response in time domain e M ( e ( r; t) p( t) A e M M m n m n m n B mn mn mn n n) ( r; t t Early-time response in Laplace domain E ( r; s) A s ( r; s) e st m m P( r; s) e ) st m N times integration A ( N ) A t Adder N Transmitted signal Tapped Delay Line N times integration t M t N A N..... AM ( N ) AM Adder Adder AMN N t N B mn A mn s n P( r; s) Scattered signal There is a duality between early-time response in the Laplace domain and late-time response in the time domain.

26 Anti-Collision Algorithm In some applications, multiple tags are present in the main beam of the antenna. In these cases, an anti-collision algorithm is needed in order to detect the number of the tags in the reader area and separate their IDs. Since the encoding process is performed based on the resonant frequencies of the structure, in the case where the tags have some similar resonant frequencies, the identification process becomes difficult. Multiple tags in the main beam of the reader antenna

27 Anti-Collision Algorithm ANT. R Central reader y R x R Local reader Center reader Chipless tag ANT. 2 R ANT. 3

28 Ranging error (cm) Cramer-Rao (CR Bound) Based on Cramer-Rao (CR) bound: ER { } c SNR C: speed of light : effective bandwidth SNR: signal to noise ratio E: error SNR (db)

29 Space-Time-Frequency Algorithm EFIE: M J ( r) J ( r r ) s e inc G ( r, r; s) J ( r r) ds tˆ E ( r) r S Current on the tags: J ( r; s) s m sm Singularity expansion: s M Nm (n) ( n) M m sm m ( n) mn N ( s sm ) m m m A J ( rr) J ( rr; s) m m UWB Antenna Tag # Tag #M Tag #2 M N m s (m) n e ( r, t) e (, t) Re e m r Rn m n ( m s ) ( t t ) m inc ( p) (, t) amp ( t Tm ) p e (t) e r m E r E r M Nm (n) s p inc stm m (, s) amps (, s) e ( n) m p nn s sm R (, ) (7) m M Nm (n) stm m Am r se ( n) m nn s sm m R

30 T (ns) E (v/m) E (v/m) T (ns) Anti-Collision Algorithm (2 different tags) x time (ns) 4 2 t 2 t frequency (GHz) k 4.9 n x -5 tag # tag #2 6 4 R Ln(R) frequency (GHz)

31 E (v/m) F (GHz) T (ns) T (ns) Anti-Collision Algorithm (2 similar tags) E 2 t 2 t frequency (GHz) Ln(R).5 x -4.5 tag # tag # frequency (GHz) R (cm)

32 T (ns) T (ns) E (mv/m) Anti-Collision Algorithm (measurement) ID: 2cm 3cm 2 3 tags 2 tags ID: time (ns) frequency (GHz) Ln( R )

33 E (v/m) F (GHz) Anti-Collision Algorithm (Space-frequency) real imaginary frequency (GHz) tags 2 tags R (cm)

34 P r /P t (db) 7.5 mm RCS (dbm 2 ) RCS (dbm ) 2 Localization of the Tags in the reader area -2 3-bit tag P r /P t (db) mm Measurement Simulation frequency Frequency (GHz) frequency Frequency (GHz)

35 Simulation and measurement set-ups TEM horn Antenna Tag z 5cm x y Simulation set-up Measurement set-up

36 (degree) (degree) (degree) Simulation and measurement Results Real position simulation measurement 4 2 Case (cm) (cm) (cm) tag in the reader area 2 tags in the reader area real position, tag real position, tag2 simulation, tag simulation, tag2 measurement, tag measurement, tag2 Case. 3 Case

37 Conclusions By introducing STMPM, we improved the accuracy of the extracted CNR from the scattered signal. Improving the resolution in both time and frequency domains. Proposing the detection, identification and localization technique based on combination of STMPM and NFMPM. Proposing the idea of collision avoidance. The proposed technique can be used in different applications.

38