Near-Field Electromagnetic Ranging (NFER) Indoor Location

Transcription

1 Near-Field Electromagnetic Ranging (NFER) Indoor Location 21 st Test Instrumentation Workshop Thursday May 11, 2017 Hans G. Schantz Q-Track Corporation Sheila Jones Naval Underseas Warfare Center Distribution Statement A Distribution: Approved for public release: distribution unlimited.

2 Agenda NFER: Near-Field Electromagnetic Ranging SOLD: Signal Optimized Location Devices Goal: sub-meter GPS-denied location accuracy 1. The Problem: Accurate Indoor Location GPS Inertial RF-Based 2. How Does NFER Work? Near-Field Wireless Project & System Description RF Fingerprinting Algorithm Locator-Receiver Beacon & Tag Transmitters 3. What Does NFER SOLD Do? Lab Testing Relevant Environments: Shoot House, Steam Tunnel, Marketplace Video Capabilities & Innovation Summary & Conclusion Contact & Acknowledgement NFER Indoor Location 2

3 The Problem: Accurate Indoor Location Global Positioning System (GPS) Gold standard for outdoor location GPS-denied environments include: Indoors Urban Canyons Underground Caves & Tunnels Inertial Navigation System (INS) High short-term accuracy Prone to drift May be confounded by user dynamics NFER Indoor Location 3

4 The Problem: Accurate Indoor Location RF-Based Location Systems Ultrawideband (UWB) High accuracy (<30cm) Line-of-Sight Worse accuracy (>1m) Cluttered Bluetooth (BLE) Low Cost Accuracy ~10-15% of range Near-Field Electromagnetic Ranging (NFER) Low frequency (~1MHz) Long Wavelength (~300m) Mean accuracy 20-40cm References: Kazimierz Siwiak, Near-Field Versus UWB Ranging, 2009 IEEE RFID Symposium, 27 April Hans G. Schantz, On the origins of RF-based location, 2011 IEEE Topical Conference on Wireless Sensors and Sensor Networks (WiSNet), February, Davide Dardari et al, Indoor Tracking: Theory, Methods, and Technologies, IEEE Transactions on Vehicular Technology, Vol. 64, No. 4, April NFER Indoor Location 4

5 How Does NFER SOLD Work? Near-Field Wireless LOW FREQUENCY ~1MHZ Part 15 Unlicensed Excellent Propagation LONG WAVELENGTH ~300M Bends Around Obstructions Multipath Resistant EXPLOIT NEAR-FIELD PHYSICS Many Parameters Very Precise Location RF FINGERPRINTING Accommodates Perturbations Yields 20-40cm Mean Accuracy HORIZONTAL POLARIZATION RADIAL POLARIZATION VERTICAL POLARIZATION NEAR-FIELD ELECTRIC MAGNETIC ELECTRIC MAGNETIC ELECTRIC MAGNETIC AMPLITUDE PHASE AMPLITUDE PHASE AMPLITUDE PHASE AMPLITUDE PHASE AMPLITUDE PHASE AMPLITUDE PHASE HORIZONTAL POLARIZATION VERTICAL POLARIZATION FAR-FIELD AMPLITUDE PHASE NO RADIAL POLARIZATION IN FAR FIELD AMPLITUDE PHASE 12 NEAR-FIELD RF PARAMETERS 4 FAR-FIELD RF PARAMETERS NFER Indoor Location 5

6 How Does NFER SOLD Work? Project & System Description NFER Signal of Opportunity Location Devices (NFER SOLD) Design, develop, and demonstrate a self-contained prototype system able to track a dismounted warfighter with high accuracy over long periods of time during T&E activities where access to GPS is unavailable or degraded. NFER SOLD yields sub-meter location accuracy in GPS-denied environment by applying RF fingerprinting algorithms to the distinctive near-field patterns of beacon transmitters. System components: Beacon Transmitters Three axes AM Band RF Magnetometer Locator-Receiver Orientation Sensors (Accelerometer, Compass) RF Fingerprinting Algorithms Locator Receiver Prototype Concept NFER Indoor Location 6

7 How Does NFER SOLD Work? RF Fingerprinting Algorithm Calibrate by logging signal data (amplitude, phase, impedance, etc) at known points. Interpolate between cal points to yield surfaces. Repeat for each available signal. Compare live data to surface values to find closest match. Least error yields predicted location. Surfaces Obtained for Four Signals Floorplan (70 x 170 ) w Cal Points Amplitudes or Received Signal Strength Indicator (RSSI) NFER Indoor Location 7

8 How Does NFER SOLD Work? Locator-Receiver Block Diagram Digital Receiver Block Diagram Locator-Receiver Phase 2 Brassboard 250ksample baseband Hilbert (not Fourier) Transforms DSP algorithms in Spartan-6 FPGA Sensors: magnetometer, accelerometer, barometer Data transmission via Lantronix xpico WiFi Module 10V antenna tuning range LPC1768 CPU Phase 2 Brassboard Enclosure 5.5 x 6.5 x 8.5 NFER Indoor Location 8

9 How Does NFER SOLD Work? Prototype Phase 3 Locator-Receiver Mechanical: Enclosure: ~9.25 x 4.25 x 1.5 Volume: ~840 cm 3 Mass: ~630 g Minimum coupling array of three 1.14in diameter ferrite antennas. Electrical: Sensitivity ~ -140dBm Channel Bandwidth ~300Hz Noise Figure < 10dB Dynamic Range: 90dB Tuning range: 1 MHz: kHz 3 MHz: kHz Power consumption: 3W + 0.7W (WiFi) Phase 3 Preliminary Mechanical Design NFER Indoor Location 9

10 How Does NFER SOLD Work? Magnetic Beacon and Compact Tag Transmitters Mechanical: Enclosure: ~10 x 8 x 7 Mass ~2.4 kg QT-641 PCB Dual 6.25 loop antennas Mechanical: Enclosure: ~2.75 x 2.25 x 1 Mass ~150g QT-641 PCB Dual 0.5 x 1 loopstick antennas Electrical: Power ~200mW Transmit Power at FCC Limit, +24dB > Tag Transmitter FCC Part Limits 1 MHz: ~ m 3 MHz: ~30 30m Tag Transmitters Electrical: Power ~200mW Transmit Power ~24dB below FCC Limit NFER Indoor Location 10

11 What Does NFER SOLD Do? Lab Testing With precision calibration, mean accuracy of 23.5 cm (9.25in) Four beacon signal sources Muscatatuck Urban Test Center Prison Steam Tunnel Shoot House Steam Tunnels Shoot House NFER Indoor Location 11

12 What Does NFER SOLD Do? Lab Testing Error Phase 3 Phase 2 Mode 20 cm 30 cm Mean 23.5 cm 30.1 cm 95% w/in 40 cm 1.40 m Phase 3 results show significant improvement over Phase 2 Phase 2 Error Distribution Phase 3 Error Distribution NFER Indoor Location 12

13 What Does NFER SOLD Do? MUTC Tunnel: Overview 190ft steam tunnel section Signal sources: 8 Beacons (1 in tunnel; 7 above tunnel) 5 Tags Signal cases: Baseline (8 Beacons + 5 Tags) Beacons Only Tags Only 4 Beacons (1494, 1503, 1521, 1557 khz) 3 Tags (1028, 1054, 1114 khz) Better than 39in (1m) accuracy: 3ft (91cm) accuracy 3 Tags (~25.9m / 85ft spacing) and Cal Points (~1.52m / 5ft spacing) 2.7ft (83cm) accuracy 3 Tags (~25.9m / 85ft spacing) and Cal Points (~3.04m / 10ft spacing) Tracking Area Side-to-Side ambiguity introduces ~5-15ft (1.5m 4.5m) errors, worst case, RX w/in inches of infrastructure 1 Mar, 2016 HW Ver. 2a SW Ver NFER Indoor Location 13

14 Steam Tunnel: NFER Indoor Location 14

15 Steam Tunnel: NFER Indoor Location 15

16 Steam Tunnel: NFER Indoor Location 16

17 What Does NFER SOLD Do? MUTC Tunnel: Beacons 1 Mar, 2016 HW Ver. 2a SW Ver NFER Indoor Location 17

18 What Does NFER SOLD Do? MUTC Tunnel: Error CDF 1 Mar, 2016 HW Ver. 2a SW Ver Summary: For ~1ft (30cm) accuracy, 5ft (1.5m) cal point density, 50ft (15m) density Beacons and/or Tags For 39in (1m) accuracy, 5ft (1.5m) cal point density, 85ft (26m) linear density Beacons and/or Tags Side-side ambiguity yields 5-15ft (1.5m 4.5m) errors, worst case Avg Error (ft) Avg Error (cm) All Signals Beacons Only Beacons Tags Only Tags NFER Indoor Location 18

19 What Does NFER SOLD Do? MUTC Shoot House: Cal Points Accuracy nominally ~30cm (1ft) Cal point density ~1.5m (~5ft) 2 Mar, 2016 HW Ver. 2a SW Ver NFER Indoor Location 19

20 Shoot House: NFER Indoor Location 20

21 Shoot House: NFER Indoor Location 21

22 Shoot House: NFER Indoor Location 22

23 What Does NFER SOLD Do? MUTC Shoot House: Tracking y (ft) Raw Data Versus Kalman Excellent accuracy from raw data Scatter ~1ft (30cm) relative to Kalman 2 Mar, 2016 HW Ver. 2a SW Ver x (ft) NFER Indoor Location 23

24 Signal sources: 8 Beacons 10 Tags Signal cases: Baseline (8 Beacons + 10 Tags) Beacons Only Tags Only 6 Outer Tags 4 Inner Tags 2 Mar, 2016 HW Ver. 2a SW Ver What Does NFER SOLD Do? MUTC Shoot House: Tracking y (ft) Various Signal Cases x (ft) NFER Indoor Location 24

25 What Does NFER SOLD Do? MUTC Shoot House: Tracking 2 Mar, 2016 HW Ver. 2a SW Ver Beacons Outer Inner Tags Only Tags Tags Only Mean Maximum Beacons Outer Inner Tags Only Tags Tags Only Error (ft) NFER Indoor Location 25

26 What Does NFER SOLD Do? Capabilities & Innovation Capabilities Accurate (<1m mean error) location in GPS-denied environments Independent of user dynamics Repeatable long-term accuracy (>12 weeks) Validated in various T&E environments including cave, tunnel, cinder block shoot house Key Innovations Orientation-independent location algorithm Compact minimum-coupling orthogonal antenna array Low noise, high sensitivity (-140dBm) receiver front end NFER Indoor Location 26

27 Summary & Conclusion Precision TSPI in GPS-denied T&E environments Low-frequency, near-field RF fingerprinting approach Accuracy <39in (1m) in a Variety of Relevant Test Environments Lab/Office MUTC Market, Tunnel, Shoot House Natural complement to inertial location: >12 week stability Locator Receiver Prototype Concept NFER Indoor Location 27

28 Contact & Acknowledgement Hans G. Schantz Sheila Jones This project is based upon work supported by the Test and Resource Management Center (TRMC) Test and Evaluation/Science & Technology (T&E/S&T) Program through the U.S. Army Program Executive Office for Simulation, Training and Instrumentation (PEO STRI) under Contract No. W900KK-13-C Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Test and Resource Management Center (TRMC) Test and Evaluation/ Science & Technology (T&E/S&T) Program and/or the U.S. Army Program Executive Office for Simulation, Training and Instrumentation (PEO STRI). NFER Indoor Location 28