High-accuracy Positioning in Multipath Channels: Location-Awareness for 5G Networks and Beyond

Size: px

Start display at page:

Download "High-accuracy Positioning in Multipath Channels: Location-Awareness for 5G Networks and Beyond"

Clemence Geraldine Stafford
6 years ago
Views:

/ Post Docs: Paul Meissner, Erik Leitinger, Stefan Grebien, Josef Kulmer, Thomas

1 1 S C I E N C E P A S S I O N T E C H N O L O G Y High-accuracy Positioning in Multipath Channels: Location-Awareness for 5G Networks and Beyond Joint work with my PhD students / Post Docs: Paul Meissner, Erik Leitinger, Stefan Grebien, Josef Kulmer, Thomas Wilding, Anh Hong Nguyen, Michael Rath , COST IRACON, Cartagena, Spain u

2 2 Introduction High-accuracy Positioning: Applications Manufacturing Retail Autonomous Driving Logistics Smart Labeling Assisted Living Objectives: Positioning and navigation; activity recognition; control Requirements: Accuracy (5 20 cm); Reliability (90 100%) Challenges: Heterogeneity: scenarios and technologies; multipath Fotos: Ubisense, SES-magotag GmbH, brighamyen.com, Witrisal, Jungheinrich, slashgear.com

3 3 Introduction Location-aware Communications: many system parameters depend on the position [Di Taranto et al., Location-Aware Communications for 5G Networks, IEEE Signal Proc. Mag., Nov. 2014]

4 High-accuracy Positioning in Dense Multipath Channels 4 Outline Four ingredients for high-accuracy positioning and location awareness Introduction Ranging and positioning in dense multipath Bandwidth scaling (1) and MIMO gain (2) Multipath-assisted indoor positioning (3) Theory and modeling Algorithms Cognitive positioning and location awareness (4) Conclusions

5 5 Ranging and positioning in dense multipath An experiment: transmission of a (UWB) pulse in an indoor environment

6 6 Ranging and positioning in dense multipath Time-of-flight Ranging

7 7 Ranging and positioning in dense multipath Time-of-flight Ranging: Problem: Multipath Radio Propagation

8 8 Ranging and positioning in dense multipath Time-of-flight Ranging: Problem: Multipath Radio Propagation

9 9 Ranging and positioning in dense multipath Time-of-flight Ranging: Problem: Multipath Radio Propagation

10 Ranging and positioning in dense multipath 10 Signal Processing for Robust Ranging (& Pos.) Experiment Modelling Performance limits Algorithms Improvements

11 11 Ranging and positioning in dense multipath Modeling the dense multipath to derive the theoretical limit (CRLB) Received signal from anchor j located at p (j) : (s(t): TX signal) [Witrisal et al. "Bandwidth Scaling and Diversity Gain for Ranging and Positioning in Dense Multipath Channels," IEEE Wireless Commun. Lett., 2016.]

12 Ranging and positioning in dense multipath 12 CRLB for ranging (in dense multipath (DM)) AWGN only AWGN + dense multipath CRLB CRLB whitening mean-squared bandwidth BW gain signal-to-noise ratio reduced SNR due to DM CRLB scales with squared bandwidth and SNR cost for nuissance estim. [Witrisal et al. "Bandwidth Scaling and Diversity Gain for Ranging and Positioning in Dense Multipath Channels," IEEE Wireless Commun. Lett., 2016.]

13 Ranging and positioning in dense multipath 13 Ranging error bound and SINR shows the bandwidth scaling bandwidth CRLB in AWGN Parameters SNR = 30 db K LOS = 1 LOS-to-DM-power CRLB in multipath detectability: ML estimator [Witrisal et al. "Bandwidth Scaling and Diversity Gain for Ranging and Positioning in Dense Multipath Channels," IEEE Wireless Commun. Lett., 2016.]

14 Ranging and positioning in dense multipath Position error bound Fisher information Position error variance (CRLB): anchor 1 J independent measurements: ranging direction matrix agent: p anchor 2

14 14 Ranging and positioning in dense multipath Position error bound Fisher information Position error variance (CRLB): anchor 1 J independent measurements: ranging direction matrix agent: p anchor 2 geometry of nodes anchor 3 [Y. Shen and M. Win, "Fundamental Limits of Wideband Localization; Part I: A General Framework," IEEE Trans. Inform. Theory, vol. 56, no. 10, pp , October ]

Ranging and positioning in dense multipath 15 (2) Diversity (MIMO) gain Fisher information J independent measurements: anchor 1 Diversity combining (SIMO, MISO, MIMO): approx.

15 Ranging and positioning in dense multipath 15 (2) Diversity (MIMO) gain Fisher information J independent measurements: anchor 1 Diversity combining (SIMO, MISO, MIMO): approx. equal geometries effective SINRs are added up: agent: p 2 x 2 MIMO anchor 2 gain = 4-fold; STDV halved SINR gain [Witrisal et al. "Bandwidth Scaling and Diversity Gain for Ranging and Positioning in Dense Multipath Channels," IEEE Wireless Commun. Lett., 2016.]

16 Ranging and positioning in dense multipath 16 Multi-antenna configurations diversity (MIMO) gain for ranging 4 x 4 MIMO vs. SISO (1 x 1) 16 independent measm. 16-fold SINR (LOS-tomultipath-power-ratio): theoretical limit reduced by factor 4 SISO ~ 1/100 1m 10cm 1cm 1/4 detection probability strongly improved 100MHz [Witrisal et al. "Bandwidth Scaling and Diversity Gain for Ranging and Positioning in Dense Multipath Channels," IEEE Wireless Commun. Lett., 2016.]

17 Ranging and positioning in dense multipath 17 CRLB for angle estimation (phased array) AWGN only AWGN + dense multipath CRLB for AoA CRLB in DM squared effective aperture signal-to-noise ratio reduced SNR due to DM CRLB scales with f c (carrier), squared aperture and SNR multipath interference limits the performance [T. Wilding, et al., "AoA and ToA accuracy for antenna arrays in dense multipath channels," in ICL-GNSS, June 2018.]

18 Ranging and positioning in dense multipath 18 Multi-antenna configurations diversity gain for angle estimation in dense multipath CRLB for AoA uniform linear array (ULA) bandwidth scales SINR matched filter diverges from CRLB minimum SINR needed for maximum likelihood 2 antenna elements 16 antenna elements 16-element array > 20-fold aperture 16-fold SINR detectability strongly improved [T. Wilding, et al., "AoA and ToA accuracy for antenna arrays in dense multipath channels," in ICL-GNSS, June 2018.]

Ranging and positioning in dense multipath 19

RFID Readers: UWB for ranging DSSS signal (@ 50 MHz) for

Communications Letters 2015] [Arnitz, Muehlmann,

19 Ranging and positioning in dense multipath 19 Application to RFID experimental validation Wideband/UWB RFID Readers: UWB for ranging DSSS signal 50 MHz) for ranging (TU Vienna, Arthaber) [Arthaber, Faseth, Galler, Communications Letters 2015] [Arnitz, Muehlmann, Witrisal; Electronics Letters 2010] [Hinteregger et al., IEEE RFID 2016]

20 20 Ranging and positioning in dense multipath Position error bound and MIMO gain: applied to multistatic RFID positioning Three configurations are compared: 1. Each reader has: 1 antenna for TX and RX (1 TRX) yielding 2 range measurements 2. Each reader has: separated TX/RX antennas yielding 2 range + 2 bistatic meas. 3. Each reader has 2 pairs of separated TX/RX ant. yielding 8 independent range plus 8 independent bistatic measurem. 114 range 4 1 range 28 bistatic 1 21 TX, TX; 1 TRX 21 RX 21 TX, TX; 1 TRX 21 RX [Hinteregger, Witrisal, et al., "MIMO Gain and Bandwidth Scaling for RFID Positioning in Dense Multipath Channels," in 2016 IEEE Intern. Conf. on RFID, ]

21 Ranging and positioning in dense multipath 21 Position error bound and MIMO gain: applied to multistatic RFID positioning Three configurations are compared: 1. Each reader has: 1 antenna for TX and RX (1 TRX) yielding 2 range measurements 2. Each reader has: separated TX/RX antennas yielding 2 range + 2 bistatic m. 3. Each reader has 2 pairs of separated TX/RX ant. yielding 8 independent range plus 8 independent bistatic measurem. [Hinteregger, Witrisal, et al., "MIMO Gain and Bandwidth Scaling for RFID Positioning in Dense Multipath Channels," in 2016 IEEE Intern. Conf. on RFID, ]

22 Ranging and positioning in dense multipath 22 Integration Measurement Campaign 3 Campaigns TU Graz Demoroom Detego Semi-industrial hall TU Wien Laboratory

23 23 Ranging and positioning in dense multipath Integration Measurement Campaign - Results TU Graz Maximum Likelihood

24 24 Ranging and positioning in dense multipath Integration Measurement Campaign - Results TU Graz Positioning Accuracy

25 Ranging and positioining in dense multipath 25 Conclusion (1) High-accuracy positioning in dense multipath requires large bandwidth and/or multi-antenna systems 5G systems to employ mm-wave and massive MIMO!

26 High-accuracy Positioning in Dense Multipath Channels 26 Outline Introduction Ranging and positioning in dense multipath Bandwidth scaling (1) and MIMO gain (2) Multipath-assisted indoor positioning (3) Theory and modeling Algorithms Cognitive positioning and location awareness (4) Conclusions

27 Multipath-assisted indoor positioning theory and modeling 27 Multipath-assisted Indoor Navigation and Tracking (MINT) concept and geometric model Idea: exploit range/position information from reflected multipath Benefits: less anchor nodes; more redundancy, i.e. robustness in NLOS; higher accuracy Geometric model (GPEM): virtual anchors (VAs) (mirror sources) [Meissner, Steiner, Witrisal, "UWB Positioning with Virtual Anchors and Floor Plan Information," in WPNC, Dresden, March 2010.]

28 Multipath-assisted indoor positioning theory and modeling 28 Signal Model (Geometry-based stochastic channel model - GSCM) Received signal: (s(t): TX signal) K deterministic multipath components Anchor (LOS), virtual anchors (NLOS), deterministic scatterers Diffuse multipath v(t) PDP MPCs characterized by

Multipath-assisted indoor positioning theory and modeling 29 Position error bound for MINT (Cramér-Rao lower bound derived from LHF) Position error variance is bounded by If no path-overlap occurs

29 Multipath-assisted indoor positioning theory and modeling 29 Position error bound for MINT (Cramér-Rao lower bound derived from LHF) Position error variance is bounded by If no path-overlap occurs (orthogonal signals from VAs) effective SINR k determines ranging information intensity for MPC from k-th virtual anchor Ranging direction matrix accounts for geometry [Leitinger, et al., "Evaluation of Position-related Information in Multipath Components for Indoor Positioning," IEEE JSAC, 2015.]

30 Multipath-assisted indoor positioning

signal model and prediction of the position

of Position-Related Information in Measured

30 30 Multipath-assisted indoor positioning theory and modeling Validation of the signal model and prediction of the position error bound [Meissner, Witrisal, "Analysis of Position-Related Information in Measured UWB Indoor Channels," EUCAP 2012.] [Meissner, "Multipath-Assisted Indoor Positioning," Ph.D. Thesis, Graz University of Technology, 2014.]

31 Multipath-assisted indoor positioning algorithms Tracking algorithms exploiting multipath (1) data association of multipath ranges and state-space tracking (2) ranging uncertainty is estimated

31 31 Multipath-assisted indoor positioning algorithms Tracking algorithms exploiting multipath (1) data association of multipath ranges and state-space tracking (2) ranging uncertainty is estimated from multipath amplitudes (3) a SLAM-style algorithm is used to discover new VAs GSCM update data assoc. GPEM update [Meissner, Witrisal, et al., "UWB for Robust Indoor Tracking: [ ], IEEE Wireless Commun. Lett., 2014.] [Leitinger, Witrisal, et al., "Multipath-assisted Indoor Simultaneous Localization and Mapping, ICC Workshops, 2015.] [Witrisal, et al., "High-Accuracy Localization for Assisted Living, IEEE Signal Processing Mag., March 2016.]

32 Multipath-assisted indoor positioning algorithms 32 Experimentation employing a lab-grade UWB channel sounder Validation steps: GSCM/GPEM-based simulation model measurement-based analysis channel-sounder-based real-time implementation (2 anchors) (1 anchor) [Meissner, et al, "Real-Time Demonstration System for Multipath-Assisted Indoor Navigation and Tracking (MINT)," IEEE ICC Workshops ]

33 33 Multipath-assisted indoor positioning algorithms Online model updates GPEM environment model GSCM channel model position and cov. of VAs SINR-values of MPCs blackboard, anchor 1 left windows, anchor 2 [Leitinger, Witrisal, et al., "Multipath-assisted Indoor Simultaneous Localization and Mapping, ICC Workshops, 2015.]

34 Multipath-assisted indoor positioning algorithms 34 Tracking performance Analysis of accuracy and robustness (non-diverging runs) Channel/environment model awareness yields robustness [Meissner, Witrisal, et al., "UWB for Robust Indoor Tracking: Weighting of Multipath Components for Efficient Estimation, IEEE Wireless Commun. Lett., 2014.]

Multipath-assisted indoor positioning - experimentation 35

Can it be replaced by low-cost hardware?

Raspberry Pi Pozyx: DW1000 + Arduino [J. Kulmer, et al.

35 Multipath-assisted indoor positioning - experimentation 35 Experimentation Lab-grade UWB channel sounder (IlmSense; > 40 keur) Can it be replaced by low-cost hardware? DecaWave DW1000 based radio nodes Decawave EVK1000 Sequitur: DW Raspberry Pi Pozyx: DW Arduino [J. Kulmer, et al., "Using DecaWave UWB Transceivers for High-accuracy Multipath-assisted Indoor Positioning," in IEEE ICC Workshops, May 2017.]

36 36 Multipath-assisted indoor positioning - experimentation DecaWave DW1000 RX Pulse shape obtained from DW1000 chip Channel 2 (500 MHz) (top); Channel 4 (900 MHz) (bottom) sampled at GHz (oversampled for time-domain fig.) unknown time alignment Raised-cosine pulse; duration ~2.4ns Raised-cosine pulse; duration ~1.5ns [J. Kulmer, et al., "Using DecaWave UWB Transceivers for High-accuracy Multipath-assisted Indoor Positioning," in IEEE ICC Workshops, May 2017.]

37 Multipath-assisted indoor positioning - experimentation 37 Analysis of Multipath-Components SINR-values in db; two alignment methods DecaWave DW1000 Channel Sounder DW1000 and channel sounder reach similar levels (difference ~ 1 2 db, consistently) SINRs: LOS is best; white board and window still promising for positioning [J. Kulmer, et al., "Using DecaWave UWB Transceivers for High-accuracy Multipath-assisted Indoor Positioning," in IEEE ICC Workshops, May 2017.]

38 38 Multipath-assisted indoor positioning - experimentation Approximate ML Positioning AWGN model (neglect DM) approximate log likelihood Monte Carlo sampling [J. Kulmer, et al., "Using DecaWave UWB Transceivers for High-accuracy Multipath-assisted Indoor Positioning," in IEEE ICC Workshops, May 2017.]

39 Multipath-assisted indoor positioning - experimentation 39 Positioning performance at 100 measurement positions; DW 1000 Ch. 4 (900 MHz) beats Ch. 2 (500 MHz) Select (virtual) anchors based on SINR values improves robustness [J. Kulmer, et al., "Using DecaWave UWB Transceivers for High-accuracy Multipath-assisted Indoor Positioning," in IEEE ICC Workshops, May 2017.]

(~ 2 GHz) to separate multipath components Solution: directive antenna; separation of multipath components in angular

40 Multipath-assisted indoor positioning - experimentation 40 Low-cost dependable UWB positioning: Using adaptable directive antenna Single-anchor positioning: Multipath components are exploited Challenge: Very large bandwidth needed (~ 2 GHz) to separate multipath components Solution: directive antenna; separation of multipath components in angular domain [M. Rath, et al., "Multipath-assisted Indoor Positioning Enabled by Directional UWB Sector Antennas," in IEEE SPAWC, 2017.]

Multipath-assisted indoor positioning - experimentation 41 Sectorized Antenna: Performance Results

Bandwidth is 500 MHz conventional antenna sectorized antenna Likelihood function: Probability of p

cm 25 cm @ 90 %) Robustness (55 % 90 % @ 25 cm) Outliers (10 % > 0.6 m 0 %) [M. Rath, et al.

41 Multipath-assisted indoor positioning - experimentation 41 Sectorized Antenna: Performance Results Experiments based on measured signals One fixed anchor at a 1 Agent positon p to be estimated Bandwidth is 500 MHz conventional antenna sectorized antenna Likelihood function: Probability of p given the measurement Sectorized antenna yields clear solution CDF of position error: Accuracy (60 cm %) Robustness (55 % cm) Outliers (10 % > 0.6 m 0 %) [M. Rath, et al., "Multipath-assisted Indoor Positioning Enabled by Directional UWB Sector Antennas," in IEEE SPAWC, 2017.]

x 7 array at the anchor (massive MIMO) AoA resolution: multimodality is reduced [Witrisal, et

42 Multipath-assisted indoor positioning algorithms 42 Maximum-Likelihood Positioning using a single mm-wave access point accurate estimation of multipath delays high position accuracy 7 x 7 array at the anchor (massive MIMO) AoA resolution: multimodality is reduced [Witrisal, et al., "High-Accuracy Localization for Assisted Living, IEEE Signal Processing Mag., March 2016.]

43 Multipath-assisted indoor positioning 43 Conclusions (2) A multipath-assisted indoor positioning system relies on models of geometry and channel characteristics (environment model) benefits from location awareness, raising robustness yields location awareness to the underlying communications system (predictability of PHY-layer performance indicators) A single access point can support high-accuracy positioning in UWB, 5G networks, etc. large bandwidth and smart antenna yield delay and angle resolution of MPCs

44 High-accuracy Positioning in Dense Multipath Channels 44 Outline Introduction Ranging and positioning in dense multipath Bandwidth scaling (1) and MIMO gain (2) Multipath-assisted indoor positioning (3) Theory and modeling Algorithms Cognitive positioning and location awareness (4) Conclusions

Cognitive positioning and location awareness 45 Cognitive dynamic system an engineering view inspired by cognitive neuroscience Perception-action cycle (PAC) sensed environment in a closed loop

45 Cognitive positioning and location awareness 45 Cognitive dynamic system an engineering view inspired by cognitive neuroscience Perception-action cycle (PAC) sensed environment in a closed loop control of sensing Cognitive perceptor extracts and separates relevant information (memory, attention) Cognitive controller act on the environment to maximize information gain Probabilistic reasoning mediates reciprocal coupling Hierarchical structure different abstraction layers [Haykin and Fuster, Proceedings of the IEEE, 2014.] [Leitinger, Cognitive Indoor Positioning and Tracking using Multipath Channel Information, Ph.D. Thesis, TU Graz, 2016.]

High-accuracy Positioning in Dense Multipath Channels 46 Conclusions Large bandwidth / multiple antennas yields high accuracy Exploiting multipath yields high robustness Cognitive positioning to

46 High-accuracy Positioning in Dense Multipath Channels 46 Conclusions Large bandwidth / multiple antennas yields high accuracy Exploiting multipath yields high robustness Cognitive positioning to yield high efficiency, low latency, etc. foreseen evolution of wireless networks 5G: mm-wave, massive MIMO is becoming available (e.g.) with 5G mm-wave systems environment model mission critical positioning massive RFID tag populations full location awareness Time has come for high-accuracy positioning, given current technology trends Location awareness is a way of the future for communications and positioning in 5G and beyond

Model Needs for High-accuracy Positioning in Multipath Channels

1 Model Needs for High-accuracy Positioning in Multipath Channels Aalborg University, Aalborg, Denmark Graz University of Technology, Graz, Austria Introduction 2 High-accuracy Positioning Manufacturing