SpotFi: Decimeter Level Localization using WiFi. Manikanta Kotaru, Kiran Joshi, Dinesh Bharadia, Sachin Katti Stanford University

SpotFi: Decimeter Level Localization using WiFi Manikanta Kotaru, Kiran Joshi, Dinesh Bharadia, Sachin Katti Stanford University

Applications of Indoor Localization 2 Targeted Location Based Advertising Indoor Navigation (e.g. Airport Terminals) Real Life Analytics (Gym, Office, etc..) Indoor localization platform providing decimeter-level accuracy could enable a host of applications

Easily Deployable Commercial WiFi chips 3

Easily Deployable Commercial WiFi chips No hardware or firmware change 4

Easily Deployable Commercial WiFi chips No hardware or firmware change No User Intervention 5

Easily Deployable, Universal Localize any WiFi device No specialized sensors 6

Easily Deployable, Universal, Accurate Error of few tens of centimeters 1 m 7

State-of-the-art System Deployable Universal Accurate RADAR, Bahl et al, 00 HORUS, Youssef et al, 05 ArrayTrack, Xiong et al, 13 PinPoint, Joshi et al, 13 CUPID, Sen et al, 13 LTEye, Kumar et al, 14 Phaser, Gjengset et al, 14 Ubicarse, Kumar et al, 14 8

State-of-the-art 9 System Deployable Universal Accurate RADAR, Bahl et al, 00 HORUS, Youssef et al, 05 ArrayTrack, Xiong et al, 13 PinPoint, Joshi et al, 13 CUPID, Sen et al, 13 LTEye, Kumar et al, 14 Phaser, Gjengset et al, 14 Ubicarse, Kumar et al, 14 SpotFi, Kotaru et al, 15

10 System Overview

Localization - Overview 11

Localization - Overview 12

Challenge - Multipath 13

Subcarriers Solving The Multipath Problem State-of-the-art SpotFi Model signal on antennas alone Model signal on both antennas and subcarriers Antennas f 1 f 2 f 3 14 f 4

θ 1 θ 2 Step 1: Resolve Multipath 15

Signal Modeling Equal Distance Line 16

Phase Phase 1 / frequency 0 Distance travelled by the WiFi signal 17

Signal Modeling AoA (Angle of Arrival) Equal Phase Line 18

Signal Modeling - AoA Define Φ = e Phase at the antenna 1: x = Γ θ 1 Phase at the antenna 2: x = Γ Φ 3 2 1 Phase at the antenna 3: x = Γ Φ 19 Γ is complex attenuation of the path. Φ depends on AoA

Say There Are Two Paths 20

Say There Are Two Paths x = Γ x x = Γ Φ = Γ Φ 21

Say There Are Two Paths x = Γ + Γ x = Γ Φ + Γ Φ x = Γ Φ + Γ Φ 22

Problem Statement x = Γ + Γ x = Γ Φ + Γ Φ x = Γ Φ + Γ Φ CSI - Known 23

Problem Statement x = Γ + Γ x = Γ Φ + Γ Φ x = Γ Φ + Γ Φ Parameters - Unknown 24

Problem Statement x = Γ + Γ x = Γ Φ + Γ Φ x = Γ Φ + Γ Φ Number of paths (or AoAs) < Number of antennas (or equations) 25

Typical Indoor Multipath 26

That s A Problem State-of-the-art Commodity WiFi chips Number of antennas/equations should be atleast 5 27

Subcarriers How To Obtain More Equations? Model signal on both antennas and subcarriers Antennas f 1 f 2 f 3 f 4 28

Each Subcarrier Gives New Equations f 2 f 1 29

Signal Modeling ToF (Time of Flight) Define Ω = e Phase at first subcarrier: x = Γ Phase at second subcarrier: x = Γ Ω 30 Γ is complex attenuation of the path. Ω depends on incoming signal ToF

Estimate both AoA and ToF More number of equations in terms of parameter of our interest 31

Say There Are Two Paths At first subcarrier, for 3 antennas x = Γ x x = Γ Φ = Γ Φ At second subcarrier, for 3 antennas y y = Γ Ω = Γ Φ Ω 32 y = Γ Φ Ω

Say There Are Two Paths At first subcarrier, for 3 antennas x = Γ + Γ x = Γ Φ + Γ Φ x = Γ Φ + Γ Φ At second subcarrier, for 3 antennas y = Γ Ω + Γ y = Γ Φ Ω + Γ Φ Ω 33 y = Γ Φ Ω + Γ Φ Ω

Subcarrier 2 Subcarrier 1 Problem Statement CSI - Known x = Γ + Γ x = Γ Φ + Γ Φ x = Γ Φ + Γ Φ y = Γ Ω + Γ y = Γ Φ Ω + Γ Φ Ω 34 y = Γ Φ Ω + Γ Φ Ω

Subcarrier 2 Subcarrier 1 Problem Statement Parameters - Unknown y = Γ + Γ y = Γ Φ + Γ Φ y = Γ Φ + Γ Φ y = Γ Ω + Γ y = Γ Φ Ω + Γ Φ Ω 35 y = Γ Φ Ω + Γ Φ Ω

Subcarrier 2 Subcarrier 1 Problem Statement x = Γ + Γ Number of equations = Number of Subcarriers x Number of Antennas x = Γ Φ + Γ Φ x = Γ Φ + Γ Φ y = Γ Ω + Γ y = Γ Φ Ω + Γ Φ Ω 36 y = Γ Φ Ω + Γ Φ Ω

AoA, ToF Estimates θ 1, τ 1 θ 2, τ 2 37

θ 1, τ 1 θ 1, τ 1 θ 2, τ 2 Step 2: Identify Direct Path 38

AoA, ToF Estimates θ 1, τ 1 θ 2, τ 2 39

Use Multiple Packets θ 1, τ 1 θ 2, τ 2 40

Use Multiple Packets 41

Use Multiple Packets 42

Use Multiple Packets 43

Direct Path Likelihood Higher weight Smaller ToF Higher weight Lower weight Lower weight Higher weight 44

Direct Path Likelihood Lower weight Higher weight Smaller ToF Tighter Cluster Lower weight Lower weight Lower weight 45

Direct Path Likelihood Lower weight Higher weight Higher weight Lower weight Smaller ToF Tighter Cluster More Packets Lower weight 46

Highest Direct Path Likelihood 47

θ 1, τ 1 θ 1, τ 1 θ 2, τ 2 Step 3: Localize The Target 48

Use Multiple APs Direct Path AoA = 45 degrees Signal Strength = 10 db Direct Path AoA = -45 degrees Signal Strength = 20 db Direct Path AoA = 10 degrees Signal Strength = 30 db 49 Find location that best explains the AoA and Signal Strength at all the APs

Use Different Weights Direct Path AoA = 45 degrees Signal Strength = 10 db Direct Path Likelihood Direct Path AoA = -45 degrees Signal Strength = 20 db Direct Path Likelihood Direct Path AoA = 10 degrees Signal Strength = 30 db Direct Path Likelihood Use different weights for different APs 50

51 Evaluation

Testbed Access point 52 52 m Target Locations AP Locations Target

Empirical CDF Indoor Office Deployment 16 m ArrayTrack Ubicarse SpotFi 0.3 m 0.4 m 0.4 m 52 m Target Locations 53 AP Locations 1 0.8 0.6 0.4 m 0.4 0.2 0 0.05 0.5 5 Localization Error (m)

Stress Test Obstacles Blocking The Direct Path 54 52 m Target Locations AP Locations

Empirical CDF Stress Test Obstacles Blocking The Direct Path 55 52 m Target Locations AP Locations 1 0.8 0.6 1.3 m 0.4 0.2 0 0.05 0.5 5 Localization Error (m)

Empirical CDF Effect of WiFi AP Deployment Density 1 0.8 0.6 0.4 0.2 3 APs 4 APs 5 APs 0.8 m 0 0.05 0.5 5 Localization Error (m) 56

Conclusion Deployable: Indoor Localization with commercial WiFi chips Accurate: Accuracy comparable to state-of-the-art localization systems which are not suitable for wide deployments Universal: Simple localization targets with only a WiFi chip 57

References J. Xiong and K. Jamieson, Arraytrack: A fine-grained indoor location system, NSDI 13. S. Kumar, S. Gil, D. Katabi, and D. Rus, Accurate indoor localization with zero start-up cost, MobiCom 14. P. Bahl and V. N. Padmanabhan, Radar: An in-building rf-based user location and tracking system, INFOCOM 2000. S. Kumar, E. Hamed, D. Katabi, and L. Erran Li, Lte radio analytics made easy and accessible, SIGCOMM 14. J. Gjengset, J. Xiong, G. McPhillips, and K. Jamieson, Phaser: Enabling phased array signal processing on commodity wifi access points, MobiCom 14. M. Youssef and A. Agrawala, The horus wlan location determination system, MobiSys 05. S. Sen, J. Lee, K.-H. Kim, and P. Congdon, Avoiding multipath to revive inbuilding wifi localization, MobiSys 13. K. Joshi, S. Hong, and S. Katti, Pinpoint: localizing interfering radios, NSDI 13. M. Kotaru, K. Joshi, D. Bharadia, S. Katti, "SpotFi: Decimeter Level Localization Using WiFi," ACM SIGCOMM 2015. All the icons are from the Noun Project https://thenounproject.com/

BackFi: High Throughput WiFi Backscatter for IoT Dinesh Bharadia *, Kiran Joshi *, Manikanta Kotaru, Sachin Katti Stanford University *co-primary authors

The Internet of Things (IoT) Vision 2

The Internet of Things (IoT) Vision Sense 2

The Internet of Things (IoT) Vision Sense Collect & Analyze 2

The Internet of Things (IoT) Vision Sense Collect & Analyze Control 2

The Internet of Things (IoT) Vision Sense Collect & Analyze Control BackFi 2

What do we need for IoT Connectivity? Sense Ubiquitous connectivity Low power High uplink rate Sufficient range 3

BackFi: Ubiquitous, low power, high throughput connectivity for IoT sensors using ambient WiFi 4

BackFi: Ubiquitous, low power, high throughput connectivity for IoT sensors using ambient WiFi Sense 4

BackFi: Ubiquitous, low power, high throughput connectivity for IoT sensors using ambient WiFi Sense Decode data 4

BackFi s Contributions Technical spec Key enabling technique Ubiquitous connectivity Low power Same as WiFi Less than 50 uw Backscatter ubiquitous ambient signals Passive backscatter radios High uplink rate Sufficient range Up to 6.67 Mbps Up to 7m Maximal ratio combining Self-interference cancelation 5

Related Work 6

Related Work Ubiquitous connectivity Low power High uplink rate Sufficient range 6

Related Work WiFi- Backscatter Ubiquitous connectivity Low power High uplink rate Sufficient range WiFi Backscatter: H. Ishizaki, et. al. A Battery-less WiFi-BER modulated data transmitter with ambient radio-wave energy harvesting B. Kellogg et. al. Wi-Fi Backscatter: Internet Connectivity for RF-Powered Devices 6

Related Work WiFi- Backscatter RFID-based Ubiquitous connectivity Low power High uplink rate Sufficient range WiFi Backscatter: H. Ishizaki, et. al. A Battery-less WiFi-BER modulated data transmitter with ambient radio-wave energy harvesting B. Kellogg et. al. Wi-Fi Backscatter: Internet Connectivity for RF-Powered Devices RFID based: J.F. Ensworth et. al. Every smart phone is a backscatter reader, P. Zhang et. al. Ekhonet 6

Related Work WiFi- Backscatter RFID-based BackFi 15 Ubiquitous connectivity Low power High uplink rate Sufficient range WiFi Backscatter: H. Ishizaki, et. al. A Battery-less WiFi-BER modulated data transmitter with ambient radio-wave energy harvesting B. Kellogg et. al. Wi-Fi Backscatter: Internet Connectivity for RF-Powered Devices RFID based: J.F. Ensworth et. al. Every smart phone is a backscatter reader, P. Zhang et. al. Ekhonet 6

BackFi s Overview Sense 7

BackFi s Overview Sense IoT Sensor 7

BackFi s Overview Sense IoT Sensor BackFi AP 7

IoT Sensor Design Sense 8

IoT Sensor Design Sense Sensor data 10101010 8

IoT Sensor Design Sense Sensor data 0 to -1 1 to +1 10101010 8

IoT Sensor Design Sense Sensor data 0 to -1 1 to +1 10101010 Data modulation 8

BackFi AP Design Sense 9

BackFi AP Design Sense Received&signal =&& Sensor& backscatter + Environmental& reflections 9

Challenge 1: Strong Environmental Reflections 10

Challenge 1: Strong Environmental Reflections 20 0 Power in dbm -20-40 -60-80 -100 40 MHz& 2.45 Frequency in GHz 10

Challenge 1: Strong Environmental Reflections 20 0 transmitted &signal Power in dbm -20-40 -60-80 -100 40 MHz& 2.45 Frequency in GHz 10

Challenge 1: Strong Environmental Reflections Power in dbm 20 0-20 -40-60 -80-100 40 MHz& 2.45 Frequency in GHz transmitted &signal sensor& backscatter 10

Challenge 1: Strong Environmental Reflections Power in dbm 20 0-20 -40-60 -80-100 Received&signal =&& 40 MHz& 2.45 Frequency in GHz Sensor& backscatter transmitted &signal sensor& backscatter environmental& reflections + Environmental& reflections 10

Why not use Self-Interference Cancelation? 11

Why not use Self-Interference Cancelation? Received signal Transmitted signal 11

Why not use Self-Interference Cancelation? Sense Received signal Transmitted signal 11

Why not use Self-Interference Cancelation? Sense Received signal Cancelation filter Σ Transmitted signal After&cancelation = 0 11

Why not use Self-Interference Cancelation? Sense Received signal Cancelation filter Σ Transmitted signal Received&signal = Sensor backscatter After&cancelation = 0 + Environmental& reflections 11

Why not use Self-Interference Cancelation? Sense Received signal Cancelation filter Σ Transmitted signal Sensor Received&signal = backscatter After&cancelation = 0 After&cancelation = 0 + Environmental& reflections 11

Eliminating environmental reflections Sense Received signal Cancelation filter Σ Transmitted signal Received&signal = Sensor backscatter After&cancelation = 0 + Environmental& reflections 12

Eliminating environmental reflections Sense Received signal Turn off the backscatter Cancelation filter Σ Transmitted signal Received&signal = Sensor backscatter After&cancelation = 0 + Environmental& reflections 12

Eliminating environmental reflections Sense Estimate the environmental reflections Received signal Turn off the backscatter Cancelation filter Σ Transmitted signal Received&signal = Sensor backscatter After&cancelation = 0 + Environmental& reflections 12

Eliminating environmental reflections Sense Estimate the environmental reflections Received signal Turn on the backscatter Cancelation filter Σ Transmitted signal After&cancelation = 0 Sensor Received&signal = + Environmental& backscatter reflections After&cancelation = Sensor&backscatter 12

Challenge 2: Inferring IoT Sensor Data!!! 13

Challenge 2: Inferring IoT Sensor Data Sense Sensor backscatter is function of:!!! 13

Challenge 2: Inferring IoT Sensor Data Sense Sensor backscatter is function of: Transmitted signal!! 13

Challenge 2: Inferring IoT Sensor Data Sense Sensor backscatter is function of: Transmitted signal IoT sensor data! 13

Challenge 2: Inferring IoT Sensor Data Sense Sensor backscatter is function of: Transmitted signal IoT sensor data Wireless channel distortions 13

Modeling Sensor Backscatter Sense 14

Modeling Sensor Backscatter Sense Transmitted signal&= : 14

Modeling Sensor Backscatter Sense : h Transmitted signal&= : 14

Modeling Sensor Backscatter : h Transmitted signal&= : 15

Modeling Sensor Backscatter : h Transmitted signal&= : Sensor data = = 15

Modeling Sensor Backscatter : h. = : h Transmitted signal&= : Sensor data = = 15

Modeling Sensor Backscatter : h. = : h Transmitted signal&= : Sensor data = = Rx=( : h. =) h Rx = sensor&backscatter = ( : h. =) h 15

Estimating Backscatter Channel : h Transmitted signal&= : 16

Estimating Backscatter Channel Use predefined sequence of sensor data&= to estimate channel : h Transmitted signal&= : 16

Estimating Backscatter Channel Use predefined sequence of sensor data&= to estimate channel : h Transmitted signal&= : Sensor data = 1 16

Estimating Backscatter Channel Use predefined sequence of sensor data&= to estimate channel : h. 1 : h Transmitted signal&= : Sensor data = 1 Rx=(: h ) h 16

Estimating Backscatter Channel Use predefined sequence of sensor data&= to estimate channel : h. 1 : h Transmitted signal&= : Sensor data = 1 Rx=(: h ) h Rx = sensor&backscatter = : (h h) 16

Estimating Backscatter Channel Use predefined sequence of sensor data&= to estimate channel : h. 1 : h Transmitted signal&= : Sensor data = 1 Rx=(: h ) h Rx = sensor&backscatter = : (h h) Estimate h 16

Modeling Sensor Backscatter : h. = : h Transmitted signal&= : Sensor data = = Rx=( : h. =) h Rx = sensor&backscatter = 17

Modeling Sensor Backscatter : h. = : h Transmitted signal&= : Sensor data = = Rx=( : h. =) h Rx = sensor&backscatter = ( : h. =) h 17

Modeling Sensor Backscatter : h. = : h Transmitted signal&= : Sensor data = = Rx=( : h. =) h Rx = sensor&backscatter = : h. =) h Incoming signal z = : h ( 17

Modeling Sensor Backscatter : h. = : h Transmitted signal&= : Sensor data = = Rx=( : h. =) h Rx = sensor&backscatter = : h. =) h Incoming signal z = : h sensor&backscatter = ( ( z. = ) h 17

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h Incoming z 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h Incoming z 1 sensor data&= -1 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h Incoming z 1 sensor data&= -1 sensor backscatter 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h Incoming z 1 sensor data&= -1 sensor backscatter WiFi AP sampling rate 40 Msps IoT sensor Information switching rate 2 Msps 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h Incoming z 1 sensor data&= -1 sensor backscatter WiFi AP sampling rate 40 Msps 20 IoT sensor Information switching rate 2 Msps 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h Incoming z 1 sensor data&= -1 sensor backscatter redundancy 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h Incoming z 1 sensor data&= -1 sensor backscatter redundancy GH:IJHK&LHMIN&ONJPIQIQR = S TUIRhMV& VHJWKUV = S T X & V X & 19

Demodulating = from Sensor Backscatter Sensor Backscatter = {E. =} h sensor backscatter redundancy GH:IJHK&LHMIN&ONJPIQIQR = S TUIRhMV& VHJWKUV = S T X & V X & 19

BackFi Prototype 24

BackFi Prototype Antenna Modulator Digital control board 24

BackFi Prototype Antenna Modulator Digital control board WiFi Backscatter radio with BPSK, QPSK & 16 PSK 24

BackFi Prototype Antenna Antenna Cancelation filter PA Modulator Digital control board TX RX WiFi Backscatter radio with BPSK, QPSK & 16 PSK 24

BackFi Prototype Antenna Antenna Cancelation filter PA Modulator Digital control board WiFi Backscatter radio with BPSK, QPSK & 16 PSK TX RX Built using WARP SDR platform, designed for 802.11, BW 20MHz, 20dBm TX power 24

BackFi Prototype 25

Testbed & Performance Metrics 6m!!! 8m!! 26

Testbed & Performance Metrics Indoor office environment: AP and IoT sensor are placed in LOS! 6m! 8m!! 26

Testbed & Performance Metrics Indoor office environment: AP and IoT sensor are placed in LOS WiFi clients are placed nearby! 8m 6m!! 26

Testbed & Performance Metrics Indoor office environment: AP and IoT sensor are placed in LOS WiFi clients are placed nearby Varied the placement of IoT device, client and WiFi AP. 8m 6m!! 26

Testbed & Performance Metrics Indoor office environment: AP and IoT sensor are placed in LOS WiFi clients are placed nearby Varied the placement of IoT device, client and WiFi AP. Performance metrics Throughput Energy per bit 8m 6m 26

What is the range and throughput? Throughput in Mbps 8 7 6 5 4 3 2 1 0 Throughput of IoT sensor with distance from AP 0.5 1 2 4 5 6 7 Range in meters 27

What is the range and throughput? Throughput in Mbps 8 7 6 5 4 3 2 1 0 Throughput of IoT sensor with distance from AP 0.5 1 2 4 5 6 7 Range in meters 100 Kbps 27

What is the range and throughput? Throughput in Mbps 8 7 6 5 4 3 2 1 0 Throughput of IoT sensor with distance from AP 0.5 1 2 4 5 6 7 Range in meters 100 Kbps Three order of magnitude better throughput than prior WiFi backscatter 27

What is the power consumption of BackFi? Throughput in Mbps EPB in pj/bit Total Power Consumption in uw for continuous mode.1 12.66 1.27.5 5.04 2.52 1 4.10 4.10 2 3.62 7.24 6.67 5.97 39.92 28

What is the power consumption of BackFi? Throughput in Mbps EPB in pj/bit Total Power Consumption in uw for continuous mode.1 12.66 1.27.5 5.04 2.52 1 4.10 4.10 2 3.62 7.24 6.67 5.97 39.92 Two order magnitude better EPB than prior work 28

Conclusion!!! 29

Conclusion BackFi provides high throughput, low power, ubiquitous connectivity using ambient WiFi signals!!! 29

Conclusion BackFi provides high throughput, low power, ubiquitous connectivity using ambient WiFi signals Not restricted to WiFi, can use other ambient signals such as LTE, Bluetooth!! 29

Conclusion BackFi provides high throughput, low power, ubiquitous connectivity using ambient WiFi signals Not restricted to WiFi, can use other ambient signals such as LTE, Bluetooth Vision: Build a pervading layer of connectivity over all amb ient communication signals! 29